The first 12 flights were performed by Atlas Centaur LV-3C vehicles, consisting of a modified Atlas D booster topped by the Centaur LH2/LOX stage.  Centaur A, B, and C variants, equipped with RL-10A-3 engines, flew during the first five missions.  Beginning with the AC-6 flight, Centaur D, with RL-10A-3-1 single-burn and RL-10A-3-3 dual-burn engines, was used. 

Atlas Centaur went on to become a workhorse for NASA, the Department of Defense, and commercial launch custormers.  Upgraded Centaur stages were also adopted for use atop Atlas IIA, Atlas III, Atlas V, Titan IIIE, and Titan IV launchers. 

Atlas Centaur Conceived 

In 1956, Krafft Ehricke of General Dynamics Convair (GD) began to study a high-energy, liquid hydrogen upper stage for Atlas.  In 1958, GD proposed and won a contract from the Advanced Research Projects Agency (ARPA) to develop Centaur.  The stage would be powered by two new 67,000 Newton turbopump fed Pratt & Whitney (P&W) RL10 engines.  ARPA provided $36 million to GD and $23 million to P&W for Centaur development. 

Centaur would weigh 13,500 kg and, when boosted by a modified Atlas, would be able to propel a 1,134 kg payload to escape velocity.  The new stage would use Atlas-style 3.05-meter diameter stainless steel balloon tanks, with a common bulkhead between an upper cylindrical LH2 tank and a lower, elliptical liquid oxygen (LOX) tank.  The bulkhead, composed of two stainless steel face sheets separated by a 6.3-millimeter layer of fiberglass insulation in vacuum, would prove to be one of several tough technical challenges. 

Tough as the technical issues were, program mismanagement would initially be a bigger problem.  The project became a joint ARPA-U.S. Air Force effort in 1958.  In 1959, NASA assumed ARPA's roll.  One year later, NASA shifted Centaur management to its new Marshall Space Flight Center (MSFC), but Werhner von Braun's team, already busy with Saturn development, gave Centaur little attention, actually stalling project funding at one point.

Still, plans called for Centaur to launch NASA Mariner flyby missions to Venus and Mars, Surveyor moon landers, and U.S. Air Force Advent geosynchronous communication satellites.  The first launch was planned for 1961 January.  Operational missions were to begin shortly thereafter.

Development problems soon delayed the program, however.  On November 7, 1960, the first dual RL-10 test at Pratt & Whitney's E-5 vertical dual-engine test stand in West Palm Beach, Florida ended in an explosion when only one of the two engines started, igniting unburned propellant flowing from the unstarted engine.  The cause of the problem was not fully determined before a second, more serious explosion occurred in January 1961 during another dual-engine test on the same test stand. The cause was finally determined to be incomplete liquid oxygen gasification at the igniter tip, but not before a third explosion occurred in a horizontal test stand during January.

Meanwhile, designers discovered that the LH2 tank common bulkhead face sheet developed cracks and leaked when exposed to the cryogenic fuel.  The sheet had to be thickened, adding weight to the stage. 
 

Test Failure

The first Atlas Centaur, variously denoted "C-1" or "F-1", was erected on the new Launch Complex 36A at Cape Canaveral during the spring of 1961.  It stood there for 15 months before finally lifting off on 1962 May 8 on a planned suborbital test flight.  The vehicle rose cleanly and steered downrange, but 49 seconds after liftoff a portion of one of four Centaur LH2 tank "weather shield" insulating panels ripped away.  The LH2 fuel tank quickly overheated, overpressurized, and failed, shredding the Centaur stage and, rapidly thereafter, destroying the entire Atlas vehicle about 55 seconds after liftoff. 

Centaur's fiberglass foam core weather shield panels were designed to keep the LH2 fuel cool, protected from aerodynamic heating effects, until shortly before Centaur separated from Atlas.  During the Atlas sustainer burn, when the vehicle had risen to an altitude of about 50 miles, well above the densest layers of atmosphere, the panels were to be jettisoned to save weight.  The shields were hinged at the bottom and held in place at the top by spring loaded tension straps and explosive bolts.   The design turned out to be structurally inadequate, however.  During the C-1 flight, a shock wave at the base of the nose cone impinged on the panels, causing negative pressure that dislodged the strap. 

The failure led to a congressional investigation by the House Subcommittee on Science and Astronautics.  The Subcommittee focused on what it called "weak" program management caused by the joint USAF/MSFC arrangement.

In 1962 August, Werhner von Braun recommended that Centaur be canceled and replaced with Saturn C-1, fitted with an Agena third stage, for the Surveyor and Mariner missions.  Meanwhile, design changes had slashed Centaur's escape payload capacity to 810 kg, causing ARPA to cancel its Advent plans.
 

New Management

In September, NASA transferred the Centaur program from MSFC to Lewis Center in Cleveland, Ohio, under the direction of Abe Silverstein.  Silverstein implemented an extensive ground test program and reconfigured the flight test plan.  The weather shield was redesigned, fuel leak problems were solved, and escape payload weight rose to 952 kg, but program costs rose from $59 million to $350 million. 

On November 27, 1963, Atlas Centaur 2 (AC-2) finally lifted off from LC36A.  The Atlas 126D booster rose cleanly, steered down a 100.5 degee flight azimuth from the Cape, and dropped its twin chamber booster package 150 seconds into the mission.  About 79 seconds later, the Atlas sustainer shut down.  Two small Atlas vernier engines burned for another five seconds before solid fuel retro-rockets at the base of the booster fired to pull Atlas away from the Centaur stage.  Centaur's two RL10A-3 engines ignited 9 seconds later to start a single 380 second burn.  For the first time, liquid hydrogen fueled rockets fired in space.  After Centaur main engine cutoff (MECO), Centaur's vernier engines fired for 12 more seconds to put the stage into 547 x 1,691 km x 30.4 degree orbit.  Centaur's weather shield panels and payload fairing were not jettisoned during the simplified mission, which included use of a crude open-loop Centaur guidance system.  The Centaur stage, not burdened with a payload, entered a high orbit where it remains today. 
 

Flight Test Failures 

Next up was AC-3 on June 30, 1964.  The flight was planned to repeat the AC-2 mission, with weather panel and payload fairing events added.  Four seconds after Centaur's RL10A-3 engines started, however, a shaft failed on one of the engine hydraulic pumps.  Without the pump, dual actuators could not gimbal the engine during the remainder of the burn.  Centaur began to roll first in one direction, then the other like a coiled spring.  The roll eventually created enough sloshing to uncover the LOX pump inlet, shutting down the engines 253 seconds into the planned 377 second burn.  Centaur failed to reach orbit, impacting in the Atlantic 4,356 km downrange, but the guidance system was able to stabilize the stage after the burn long enough to allow a test of four, 23 kgf thrust hydrogen peroxide ullage motors.  The burn settled the propellants in their tanks despite weightless conditions, allowing the RL10A-3 boost pump to restart, a prelude to restarting the engines in space.  AC-3 successfully jettisoned its weather panels and payload fairing, and Centaur's Honeywell guidance system was successfully test flown, though out of the control loop, for the first time.

AC-4, launched December 11, 1964, was to perform the first RL10A-3 restart in space.  Centaur, which carried a 952 kg pound dummy Surveyor payload, performed a successful first burn to put itself into a 165 x 178 km x 30.7 degree parking orbit.  After a 25 minute coast, however, Centaur's engines failed to restart, preventing the stage from reaching its planned 160 x 8,000 km final orbit.  Engineers had downsized Centaur's four hydrogen peroxide ullage rockets from the successful AC-2 design.  The new thrusters provided less than 0.5 kgf thrust each in a misguided effort to reduce fuel mass.  The new verniers turned out to be insufficient to settle propellants during the coast period.  As a result, LH2 venting put Centaur into a tumble-roll, uncovering the propellant inlets.  Since NASA listed the RL10 restart as a secondary objective, and since Centaur had successfully flown using closed loop Honeywell guidance for the first time, AC-4 was reported to be a success at the time.  Today, the result would clearly be called a failure. 

The low point of Atlas Centaur development may have occurred on March 3, 1965, when AC-5, on a planned single burn Centaur mission carrying the Surveyor SD1 dynamic mass model, exploded on Launch Complex 36A.  Two seconds after liftoff, the Atlas 156D RP-1 fuel pre-valves for the Rocketdyne YRL-89NA-7 booster engines suddenly closed.  Fuel pre-valves are normally only used during pre-launch activities.  Both 70,750 kgf thrust engines shut down, causing AC-5 to fall back through the launch stand and explode in the biggest fireball ever seen on a Cape launch pad. 

LC36A was badly damaged.  Making matters worse was the fact that NASA had shortsightedly mothballed nearby LC36B - after finishing 90% of its construction - to save money.  In the wake of the accident, NASA decided to simultaneously finish LC36B and rebuild LC36A. 

Success at Last

LC36B was completed in record time.  On August 11, 1965, AC-6 lifted off from it, carrying Surveyor SD2 on a single burn Centaur mission.  For the first time in nearly two years, Atlas Centaur scored a complete success.  Centaur burned for 435 seconds to boost SD2 into a 166 x 815,085 km x 28.6 degree orbit with a 31 day period and an apogee more than twice as far from Earth as the Moon.  An upgraded Atlas, equipped with engines that provided 176,417 kgf thrust versus the previous 166,440 kgf, boosted AC-6.

The flight proved that Atlas Centaur was finally ready to perform the direct ascent Surveyor mission.  Now Surveyor itself, delayed by mass-reducing design changes, became the pacing item.  During late 1965 and early 1966, NASA and GD prepared the AC-7 and AC-10 vehicles for direct ascent Surveyor missions using RL10A-3-1 Centaur engines.  Simultaneously, the AC-8 and AC-9 vehicles were prepared for additional dual-burn Centaur test flights using RL10A-3-3 engines and refitted with four 23 kgf thrust ullage motors.

AC-8 lifted off first from LC36B on April 7, 1966, carrying Surveyor SD3.  The RL10A-3-3 engines successfully performed a 325 second first burn to put Centaur into a 175 x 344 km x 30.7 km parking orbit.  Sadly, the planned 107 second Centaur second burn failed at the start when the more powerful ullage motors ran out of hydrogen peroxide fuel just before the end of the 25 minute parking orbit coast period.  Had the burn succeeded, SD3 would have been boosted to a 167 x 380,000 km orbit.  Had AC-4 carried the original ullage motors, the ullage fuel problem would have been identified 17 months earlier.

To the Moon

On May 30, 1966, AC-10 erased Atlas Centaur development frustrations when it successfully launched Surveyor 1 toward the Moon from LC36A.  The rocket rose into a clear late morning sky, framed by the launch pad's bright red mobile service tower.  Atlas booster cutoff occurred at T+142 seconds.  Sustainer cutoff came 240 seconds into the flight.  About 13 seconds later, Centaur's twin RL10A-3-1 engines begin their single, 437 second burn to boost the vehicle into a 169 x 609,333 km x 30 deg orbit.  The two-part payload shroud jettisoned during the Atlas sustainer phase, at T+203 seconds.  Surveyor 1 extended its legs 26 seconds after Centaur shutdown and separated from the stage at the T+757 second mark.  After only one mid-course correction en route, Surveyor 1 performed the first soft-landing on the Moon, in the Ocean of Storms, on June 1.

Now, at long last, Atlas Centaur hit its stride, scoring a series of consecutive successes.  AC-7 boosted Surveyor 2 on a direct ascent flight from LC36A on September 20, 1966.  The launch was successful, but Surveyor 2 crashed on the Moon on September 23.  AC-9 scored an important success on October 26, 1966 when it performed the long-awaited first in-space liquid hydrogen engine restart, boosting the Surveyor SD4 mass simulator into orbit.  The final two LV-3C Atlas Centaurs flew in 1967.  AC-12 successfully boosted Surveyor 3 to the moon on a dual-burn Centaur mission on April 16.  AC-11 launched Surveyor 4 on the last direct ascent, single burn mission on July 13.  Again, Atlas Centaur succeeded, but Surveyor 4 disappeared on July 16, probably when its Thiokol solid fuel retro-rocket exploded at startup.

AC-12 was the last LV-3C Atlas Centaur.  Fifty six subsequent Atlas Centaur missions were performed by the more powerful SLV-3C, SLV-3D, and Atlas G/Centaur D1AR variants, which served NASA until 1989.  NASA planned to phase out Atlas Centaur and other ELVs in favor of the space shuttle, and one of the LC36 pads was actually mothballed during the early 1980s, but the 1986 Challenger accident altered those plans.  General Dynamics (GD) and its succesors were allowed to market commercial launch services for the first time.  GD reopened its mothballed Atlas Centaur production line at San Diego (subsequently moved to Denver by Lockheed Martin) to build legacy "Atlas 1" models along with heavily upgraded Atlas 2, 2A, and 2AS models.

Fifty five Atlas Centaurs flew during the 1990s, far exceeding NASAs launch rate during any previous decade.  Nineteen more were launched between 2000 and 2004, but at the end of 2004 Lockheed Martin phased out the Rocketdyne-powered Atlas in favor of the new Atlas 5 Evolved Expendable Launch Vehicle series.  A total of 142 Rocketdyne-powered Atlas Centaurs were launched during the 42 year-long program. A total of 127 were successful, including all 63 examples of the final Atlas 2 and 2AS design.

 

Atlas LV-3C Centaur R&D Firing History

Mission/      Date    Site              Comments
Atlas No.
=================================================================

F-1/104D     5-8-62  CC36A   Failed suborbital flight.  Centaur 
                             weather shield failed T+49s.  
                             Exploded T+55s.

AC-2/126D  11-27-63  CC36A   Success.  Single burn Centaur flight
                             to 340 x 1050 mi x 30.4 deg orbit. 
                             No payload.  No fairing or weather
                             panel separation planned.  First 
                             successful LH2 fueled space flight. 

AC-3/135D   6-30-64  CC36A   Failed to orbit.  Broken RL-10
                             actuator hydraulic pump shaft prevented
                             one of two RL-10 engines from vectoring 
                             during planned single Centaur burn.  
                             Stage rolled, uncovering LOX pump inlet
                             and ending burn after 253 seconds, 
                             124 seconds early.

AC-4/146D  12-11-64  CC36A   First Centaur restart attempt failed, 
                             leaving 2,100 lb Surveyor mass model 
                             in 165 x 178 km x 30.7 deg orbit, short 
                             of planned 160 x 8000 km.  Ullage 
                             rocket design flaw left pump inlets 
                             uncovered after 25 minute coast period.

AC-5/156D    3-2-65  CC36A   Failed to orbit.  Exploded on pad at 
                             liftoff when Atlas RP-1 prevalve 
                             closed at T+2s.  Planned single burn 
                             Centaur mission with 1,400 lb 
                             Surveyor SD-1 model.  LC36A damaged.   

AC-6/151D   8-11-65  CC36B   Flawless single burn Centaur flight 
                             put 950 kg Surveyor SD-2 in 
                             102 x 510,861 mi x 28.6 deg orbit 

AC-8/184D    4-7-66  CC36B   Second Centaur restart attempt failed, 
                             Leaving 950 kg Surveyor SD-3 in 
                             175 x 344 km x 30.7 deg orbit, far 
                             short of planned 167 x 380,000 km.  
                             Redesigned ullage motors ran out of 
                             fuel just before RL-10 restart. 
                             Engines shut down right after starting.

AC-10/290D  5-30-66  CC36A   Perfect single burn Centaur mission 
                             sent 950 kg Surveyor 1 toward lunar
                             landing.

AC-7/194D   9-20-66  CC36A   Perfect single burn Centaur mission 
                             sent 950 kg Surveyor 2 toward failed
                             lunar landing.

AC-9/174D  10-26-66  CC36B   First successful two-burn Centaur
                             mission.  First LH2 restart.
                             RL-10A-3-3 restarted after 25 min 
                             parking orbit coast, putting 950 kg 
                             Surveyor SD-4 into 105 x 289,000 mi 
                             orbit.  Centaur compensated for 
                             5 second early Atlas shutdown.
    
AC-12/292D  4-17-67  CC36B   Perfect two-burn Centaur mission sent 
                             950 kg Surveyor 3 toward lunar landing.
    
AC-11/291D  7-14-67  CC36A   Perfect single burn Centaur mission 
                             sent 950 kg Surveyor 4 toward failed
                             lunar landing.
==================================================================
    
CC = Cape Canaveral