while an altitude of 90,000 feet was reached. The flight ended when a hydraulic pump burned out. Subsequent investigation indicated it had been run unpressurized several times during ground tests, which damaged it.
The two successful launches sparked renewed interest, and an order for a second batch of fifteen D-21s was issued on April 29. In May, Johnson formally proposed the new launch profile. Using a B-52H as the launch aircraft, he told the air force, would improve safety, cut costs, and extend the deployment range over the short-range M- 21.
The third D-21 launch was made by Park and Beswick on June 16. Article 505 flew 1,550 nautical miles and made eight programmed turns to photograph the Channel Islands, San Clemente Island, and Santa Catalina.
The flight was perfect until hatch separation. Due to an electrical problem, this did not occur.[314]
The fourth D-21 free flight was set for July 30, 1966. The D-21, Article 504, would carry a full fuel load for the first time. This meant it was heavier than on any of the previous launches. The launch would be made at a slightly higher speed and at exactly 1.00 g. Park, as before, was flying the launch aircraft, Article 135. His launch control officer was Ray Torick.
Article 134 was used as the chase plane for the launch. It was flown by Art Peterson, with Keith Beswick in the backseat to film the separation. The chase M-21 flew about three hundred feet to the right and about one hundred feet behind the launch M-21, at a speed of more than Mach 3.[315]
As the two planes flew in formation, Park reached the launch speed of Mach 3.25, began the shallow dive, and Torick started the separation sequence. From analysis of the data, it appears that Article 504 climbed more slowly through the shock wave and suffered an unstart. This caused it to strike the back of Article 135. The M-21 pitched up and the aerodynamic loads tore off the forward fuselage. Park and Torick were subjected to in-credible g forces as the fuselage tumbled. The cockpits depressurized and Park and Torick's suits inflated. The two crewmen ejected and landed in the ocean 150 miles offshore. Park was picked up by a helicopter, but Torick, having survived a Mach 3 breakup and ejection, drowned when sea water entered his pressure suit.
Kelly Johnson was devastated by the death of Torick and personally canceled the D-21/M-21 program.[316] He had long feared launch problems and was unwilling to see any more pilots killed. He concluded that the Mach 3 launch of so large an aircraft could not be justified from a safety point of view.[317] A number of D-21s had already been produced, and rather than scrapping the whole effort, Johnson again proposed to the air force that they be launched from a B-52H. This, however, would require major modifications to the D-21. It would take a year to complete the work.
Using the B-52H as drop plane entailed a complete rebuilding of the D-21s.
The process involved removing the outer wing panels, inlet cone, and ramjet from the airframe. Attachment points were added to the top of the fuselage for the pylon and to the bottom for the rocket booster. Once this was completed, the drone would be reassembled and its systems checked out.
Due to the major changes made, the drone was redesignated the 'D-21B.'
The program code name remained Tagboard. Conversion work was under way in Burbank in late 1966 and early 1967.
The rocket booster used to propel the D-21B to ramjet ignition speed was longer and heavier than the D-21 itself. The booster was 44.25 feet long, had a diameter of 30.16 inches, and weighed 13,286 pounds. It was cylindrical with several ridges, giving it the appearance of a water pipe. At the pointed nose was the propeller of a ram air turbine, which spun to provide electrical power. The solid rocket motor produced an average thrust of 27,300 pounds and burned for 87 seconds. To stabilize the assembly during the burn, a fin was attached to the bottom of the booster. To provide ground clearance, it folded to the right while attached to the B-52. The total weight of the D-21B and its booster was over 24,000 pounds.
Two B-52Hs were modified to act as launch planes. The major modification was the addition of two large pylons to hold the D-21B. These were much larger than the pylons for the Hound Dog cruise missile normally carried by B-52Hs. They bolted to the existing attachment points and involved no changes to the wings' structure. Inside the two B-52s, two LCO stations were added to the rear of the flight deck; each station was independent, with its own command and telemetry system, as well as a periscope.
The command system allowed the LCO to activate postlaunch functions normally operated by the drone's programming (engine ignition, booster jettison, telemetry, hatch ejection, and destruct). This provided a backup should the programming fail, or if it was necessary to change the timing.
The telemetry system recorded data on the functioning of the flight control, propulsion, fuel, booster, electrical and hydraulic systems, engine and equipment temperatures, as well as the D-21B's Mach number, direction, and location. This information was used to monitor the launch, and for postflight analysis of any problems. The command and the telemetry systems were duplicated for reliability. A stellar navigation system was also added to the B-52. This was used to update the D-21B's own inertial navigation system during the long flight to the drop point.
Finally, an air-conditioning system provided air to the D-21B for temperature control and to drive the auxiliary power unit (APU).[318] Temperature control was critical, as the D-21B would be 'cold soaked' by the negative-58 degree F conditions during the long flight to the launch point. After the drop, the D-21 would be suddenly heated by the acceleration to Mach 3-plus cruise. This put severe thermodynamic stresses on the vehicle and its systems.[319]
The first of the B-52s arrived at Palmdale on December 12, 1966, to begin the modifications. The 4200th Test Wing at Beale Air Force Base was assigned to undertake both the test launches and the operational missions.[320]
Before launch, the LCO would lower the booster fin, turn on the telemetry, test the automatic flight control system, and turn on the fuel and the observation camera. The drop would be made at about 38,000 feet. The assembly would fall free for a moment, then the booster would ignite. It would accelerate forward, then pitch up into the climb. At 50,000 feet, it would go into the final climb trajectory. As it passed through 63,000 feet, the destruct system would be activated; the LCO could destroy the vehicle should it go off-course. At 74,000 feet, the ramjet would ignite.
Soon after, the APU would take over the electrical load. When the D-21B reached an altitude of about 80,000 feet, two explosive bolts would fire and the booster would drop off. The D-21B's automatic flight control system would then go to a preprogrammed Mach number. During the first ten minutes of the cruise, the LCO could send a destruct command to the D-21B.
After this point, the telemetry and command system was turned off.
The D-21B would be controlled during the overflight by an onboard inertial navigation system. This was preprogrammed with the headings, bank angles, and the camera on-and-off points. Should there be a problem, the D-21B would automatically destruct. This would be triggered by any loss of altitude.
When the mission was completed, the D-21 would fly to friendly airspace. Once clear, the command system, recovery beacons, and telemetry would be turned back on. The D-21B could now be controlled by the JC-130 recovery aircraft. Over the recovery zone, the ramjet's fuel supply would be cut off. The D-21B would slow, then go into a dive. As it passed through 60,000 feet, at a speed of Mach 1.67, the latches on the front edge of the hatch would release. It would open like a door and separate from the D-21B's airframe. The recovery parachute would open, and the hatch would be caught in midair by the JC-130. (The procedure was similar to that used to recover reconnaissance satellite capsules.) If the midair recovery was unsuccessful, the hatch would float, and could be picked up by a ship. The D-21B would continue its descent, until it self-destructed at 52,000 feet.[321]
Support for the Tagboard program was starting to grow. On January 18, 1967, Kelly Johnson had a meeting with Deputy Secretary of Defense Cyrus Vance. Vance said he was very much for the project and asked Lockheed to press forward with it. He said that the U.S. government would never again fly a manned aircraft over enemy territory in peacetime.
By late summer of 1967, the modification work to both the D-21Bs and the B-52Hs was complete. The test program could now resume. In effect, the development program would have to begin again from scratch. The three successful D-21 flights had provided limited data. They had flown only half the D-21's full range, and the hatch had
