Thunder Lost - The Delta 3 Story
Thirteenth in a Series Reviewing Thor Family History
by Ed Kyle, Updated 8/29/2010

The Search for More Payload

Delta 269, the Second Delta 3, at Pad 17B in 1999

While assembling its 1986 bid for the U.S. Air Force Medium Launch Vehicle (MLV) program, McDonnell Douglas considered using Japan's then in-development H-1 liquid hydrogen upper stage and/or its LE-5 engine.   Discussions were held with Japanese officials about the possibility of creating a "joint" launch vehicle that would marry the U.S. Delta 3920 first stage to Mitsubishi's second stage, presumably to fly from both countries.  In the end the idea was bypassed in favor of upgrading the existing Delta design to create the MLV winning "Delta 2".     

In 1988, the U.S. Air Force offered another launch vehicle program for bids.  The "MLV II" program's specifications (2.6 tonnes to GTO) seemed to align with the capabilities of General Dynamics Atlas Centaur, but McDonnell Douglas nonetheless offered a bid.  Its design used a liquid hydrogen/liquid oxygen upper stage, to be built by Martin Marietta, married to the then-in-development Delta 2 first stage and strap-on motors.  General Dynamics won the competition in May 1988 with its improved Atlas Centaur, which it named "Atlas 2".   

After Delta 2 entered service in 1989-90, McDonnell Douglas again considered a high energy upper stage.  Delta 2 could lift more than 1.8 tonnes to GTO, enough to handle the then-common Hughes HS-376 spin-stabilized satellites, but spacecraft were growing heavier.  In 1987, Hughes began offering its 3-axis stabilized HS-601, which would weigh 2.5 to 3.1 tonnes.  The satellite would require a 4 meter diameter payload fairing, substantially wider than Delta 2's 2.9 meter fairing. 

Designers initially contemplated a "Delta 7930" design similar to the MLV II bid.  It was a Delta 2 with a 3.2 meter diameter liquid hydrogen second stage powered by a single RL10 engine.  This rocket would have been able to lift 2.6 tonnes to GTO.  It also would have adhered to Delta's long-term incremental growth tradition.
   

Delta 3 Materializes

Delta 3 (Delta 280) Cutaway Drawing (Boeing)

In the post-Challenger commercial launch era, incremental growth wasn't enough.  Hughes even-heavier HS-601HP grew to more than 3.5 tonnes.   McDonnell Douglas had to switch to a wide-body, 4 meter diameter second stage, and more-powerful Alliant 46 inch diameter Graphite Epoxy Motors (GEM-46) strap-on solids to lift the heavier stage and payload.   The first stage was also modified with a shorter but fatter kerosene fuel tank (4 meters rather than 2.4 meters diameter), to allow the new rocket to fit within the Delta 2 service tower. 

On May 10, 1995, McDonnell Douglas announced that it would develop the new rocket, which it named "Delta 3", using more than $200 million of its own funds, with a planned first launch in 1998.  Payload capacity to GTO had grown to 3.8 tonnes, with nearly 8.3 tonnes possible to LEO from Cape Canaveral. 

On the same day, Hughes and McDonnell Douglas announced that they had signed a contract for 10 Delta 3 launches.  The companies increased the total to 13 launch contracts in subsequent months, creating the possibility of a $1.5 billion total contract value.  By 1998, Hughes had assigned three of the launches to the NASA/NOAA GOES N, O, and P weather satellites and five to ICO Global Communications for its Medium Earth Orbit constellation.  In 1997, Space Systems/Loral reserved five Delta 3 launches, increasing the Delta 3 backlog to 18 launches through 2002.

International Competition, and Merger

A factor that came into play during the mid-1990s, with the end of the Cold War, was Russia's push to increase its commercial space launch quota, specifically for the Proton rocket.  McDonnell Douglas opposed increasing the Russian quota in 1996, as it had opposed an earlier Ukrainian quota for the Zenit rocket.  The company specifically noted that higher foreign quotas would cause it to lose value in its Delta 3 investment. 

Meanwhile, satellite manufacturers such as Hughes called for eliminating the quotas to reduce their launch costs.  At the time, U.S. satellite makers held a much larger worldwide commercial market share than U.S. launch providers.  In the end, the satellite argument won.

Another post-Cold War reality was rationalization of the U.S. aerospace industry.  On December 15, 1996, McDonnell Douglas and Boeing announced their intention to merge under the Boeing name.  The merger was not consummated until July 1, 1997.  Although it was not immediately apparent, the merger would have decisive consequences for the Delta 3 program.

Delta 3 Design

Delta 3 First Stage Manufacturing at Pueblo, Colorado

Delta 3 was an 8930 model under the four-digit numbering system. The rocket used an 8000-series Thor first stage with nine GEM-46 strap on motors, a Rocketdyne RS-27A main engine, and a 4 meter diameter kerosene tank mounted on top of the standard 2.4 meter diameter liquid oxygen (LOX) tank.  The third digit in the model number, a "3", signified the presence of the new Delta Cryogenic Upper (second) Stage (DCUS). 

Alliant built the GEM-46 motors in Bacchus, Utah.  GEM-46 was a scaled up version of the GEM-40 used by Delta 2.  The graphite-epoxy case motors used HTPB solid propellant. Six ground-lit GEM-46 produced more than 62 tonnes of thrust each at liftoff.  Three air-lit solids produced 64 tonnes of thrust each.  At liftoff, the boosters and RS-27A main engine together created 462.7 tonnes - more than 1 million pounds - of thrust.  

Like Delta 2, Delta 3 was assembled in Pueblo, Colorado.   Huntington Beach fabricated the first and second stage liquid oxygen tanks, the interstage, and other parts.  Rocketdyne built RS-27A propulsion systems in Canoga Park, California for both rockets.. Japan's Mitsubishi Heavy Industries manufactured the 4 meter diameter second stage liquid hydrogen tank and first stage kerosene tank.  Mitsubishi used tank tooling from its H-2 stage.  H-2, a 4 meter diameter liquid hydrogen two-stage rocket, entered service in 1994.

Delta Cryogenic Upper Stage Assembly at Pueblo, Colorado

McDonnell Douglas selected a two-tank, non-common bulkhead design for the DCUS.  It was the first entirely new high energy upper stage developed in the US since the 1960s.  The company fabricated its own 3-ish meter diameter ellipsoidal liquid oxygen tank that was attached to, and hung beneath, the cylindrical 4 meter liquid hydrogen tank.  Trusses arranged in a triangular configuration provided the intertank attachment.   Five helium pressurant bottles were mounted in the intertank space.

Pratt & Whitney's existing RL10 engine design was insufficient for Delta 3.  The company offered a new RL10B-2 that produced more thrust (11.22 tonnes) and that used a big three-piece extendible carbon-carbon composite nozzle to improve specific impulse.  The nozzle was 2.5 meters long and 2.1 meters in diameter.  At 462 seconds specific impulse, the engine would be the world's most efficient. 

The RL10B-2 engine fit to the base of the liquid oxygen tank.    Another set of triangular trusses supported a doughnut shaped avionics shelf below the liquid oxygen tank, surrounding the top of the RL10B-2.  The shelf held Delta's RIFCA (Redundant Inertial Flight Control Assembly) which controlled the vehicle during all stages of flight.  Four hydrazine thruster modules, and the hydrazine bottle that provided their propellant, were also mounted to the shelf.  These provided roll control during the main engine burns and three-axis control during coast periods. 

DCUS was 10.98 meters long and 4 meters in diameter.  The RL10B-2 engine, the avionics shelf, and the LOX tank fit inside the rocket's interstage while the load-bearing LH2 tank sat above and atop the interstage, its orange spray-on insulation visible to the outside world.   The stage could restart up to two times, with a total burn time of up to 700 seconds. 

Delta 3's 4.0 meter diameter composite payload fairing separated into two halves in flight.  For the first time with Delta, the fairing would be used to encapsulate payloads at their processing facility before transportation to the launch pad for integration with the vehicle.  The new "encapsulation" method provided better protection for the payloads on the launch pad and reduced the launch vehicle's time on the pad by about one week compared to Delta 2. 

Huntington Beach manufactured the Delta 3 composite payload fairing and interstage.  In 1998, Boeing subcontracted similar work for Delta 4 to ATK in Luka, Mississippi.  The 1998 contract was awarded to allow components to be barged up the Tennessee River from Luka to Boeing's new Delta 4 factory in Decatur, Alabama.

Daniel Collins, who would subsequently be vice president of Boeing's Launch Systems Division and Chief Operating Officer of United Launch Alliance, served as Boeing's Delta 3 program manager.

Testing and Development

RL10B-2 with Nozzle Extensions Being Tested in Arnold Test Cell (Note Cherry-Red Carbon-Carbon Sections)

In August 1995, Pratt & Whitney awarded the RL10B-2 extendible nozzle subcontract to Europe's Snecma.  Snecma delivered its first development nozzle one year later, during August 1996. 

The RL10B-2 qualification test program included 188 tests on several engines for a total of 17,288 sec of accumulated test time. Pratt & Whitney delivered an early RL10B-2 test engine to the Arnold Engineering Development Center during August 1996, where it was mated to the Snecma nozzle extension.  In 1997, the engine was tested in Test Cell J-4 under near vacuum conditions.  

An entire Delta 3 test article second stage, called the "X-Stage", was tested in the Spacecraft Propulsion Research Facility's B-2 vacuum test chamber at NASA's Lewis Research Center's Plum Brook Station beginning in February 1998.  The testing, which lasted until early April 1998, included tanking and detanking and 13 engine firings that simulated one, two, and three-burn mission profiles for a total of 860 seconds.  The heavily instrumented X-Stage was not equipped with the extendible nozzle, because the B-2 stand could not accommodate a nozzle of that size, but was otherwise largely representative of a flight stage.
 

"X-Stage" Being Lowered into Plum Brook Test Chamber.

A second Delta 3 upper stage was tested at Goddard Space Center's Acoustic Facility in February and March of 1998.  During this testing, the stage was subjected to acoustic loading representative of launch conditions.  It was also exposed to shock loads that represented stage separation.

Only one launch site, Cape Canaveral Space Launch Complex 17B, was modified to handle Delta III. During 1997, the launch deck at 17B was strengthened and a new exhaust deflector and exhaust ducting system was added to handle the higher thrust GEM-46 solid booster motors.  New LH2 and LOX piping was added to the Umbilical Mast to support the new cryogenic upper stage.  A new LH2 storage area was constructed on the complex, southeast of the pad.

In 1998, after a January Delta 2 launch christened the rebuilt pad, a series of "pathfinder" propellant loading tests were performed to test the modified launch pad equipment for Delta 3.  

But even as the first launch campaign prepared for kick off, Delta 3 was being outmoded.  Up the coast a few miles, Boeing, now firmly in charge of former McDonnell Douglas launch systems, began work on an entirely new launch complex for the newly won Delta 4 Evolved Expendable Launch Vehicle program.  The new launch pad was expected to cost about $250 million, rivaling the entire development cost of Delta 3.

Flight

The late 1990s were not kind to rockets flown from Cape Canaveral.  Delta 241, an ultra-reliable Delta 2, exploded seven seconds after liftoff from Complex 17 on January 17, 1997.  A Titan 4 pitched over and exploded 41 seconds after launch from Complex 41 on August 12, 1998.  Two more Titan 4 upper stage failures would occur during April 1999 following launches from the Cape.  Combined, these four military satellite launch failures cost U.S. taxpayers more than $3 billion. 

In the midst of this failure string, Boeing employees, in June 1998, stacked their company's first Delta 3 on Complex 17B for a planned July launch, but the flight was delayed a few weeks as officials dealt with pyro component testing issues.  Finally given the "go", crews stacked the payload, a $225 million communications satellite named Galaxy 10.  The mission was identified as "Delta 259". 
   

Delta 259

Delta 259 During Service Tower Rollback

Delta 259 lifted off at 01:17 UTC on August 27, 1998.  The first stage RS-27A main and vernier engines, augmented by six ground-lit solid rocket motors, drove the 301 tonne rocket skyward on a combined total of more than 462 tonnes of thrust.  The rocket quickly shot upward into the night sky and turned downrange over the Atlantic.  It passed safely through "Max-Q", the region of maximum dynamic pressure when the largest forces of the flight were imparted onto the vehicle.  It flew on, nearing 20 km in altitude and 1,100 meters/second velocity, until its six ground lit motors were nearly finished with their 75 second-long burns.   

Then, suddenly, inexplicably, the ascending rocket turned sideways and was instantly replaced by an expanding fireball.  Streams of tumbling, burning debris dropped toward the ocean.   Observers watched one large object, which turned out to be Galaxy 10,  explode when it impacted the Atlantic about 15-20 km offshore. 

An investigation quickly determined that the rocket had begun to suffer 4 Hertz roll oscillations about 50 seconds after liftoff, when the RIFCA system loaded a new set of flight control system gain constants - something that it did every 10 to 15 seconds as the rocket flew through differing phases of flight.  The roll oscillation was created when the RS27A and three of the ground-lit solid motors vectored in a way that amplified a natural 4 Hertz resonance of the vehicle.  As the system fought the resonance, it rapidly used up its hydraulic fluid.  At T+65 seconds, the fluid ran out, leaving the solid motor nozzles stuck in their final position.   The RS-27A system tried to correct, but was unable to counteract the forces created by the powerful solids.  Delta 259 began to pitch over at T+72 seconds, and break apart, its final destruction ensured by a flight termination system. 

The Delta 3 guidance system, it turned out, had not been designed to handle the particular 4 Hertz roll mode that led to the failure.  During Delta 3 development, McDonnell Douglas/Boeing designers had identified 56 roll modes.   Extensive Delta 2 flight data had shown that the most significant roll mode at liftoff remained dominant throughout the first stage flight.  The team assumed, incorrectly, that Delta 3 would behave in a similar fashion.  Since the 4 Hertz roll mode was not significant at Delta 3 liftoff, the designers had not added its effects into the control system. 

But Delta 3 used substantially heavier, more powerful solid motors than Delta 2.   While three of the Delta 3 solids used thrust vector control, none of the Delta 2 solids provided steering.  It turned out that Delta 3's 4 Hertz roll mode did become more significant as the solid motor propellant was consumed. 

The relatively straightforward fix was to modify the flight control software.   Analysts noted, in hindsight of course, that full vehicle dynamic testing or more extensive flight control analyses would have discovered the roll oscillation before the launch.  Some also wondered why the inaugural flight had carried a costly live payload.  Boeing blamed the design flaw, in part, on lack of communication between design groups. 

Delta 269

Delta 269 Liftoff.  Modified SLC 17B Exhaust Duct Created Distinctive Plumes Around Rockets when Solids Ignited.

The second Delta 3 mission, Delta 269, launched with the $145 million Orion 3 communications satellite on May 5, 1999.  This time the flight proceeded flawlessly through the first stage burn, allowing the RL10B-2 second stage engine to perform its first flight burn, pushing the stage and payload into a parking orbit. 

The stage coasted until about 21 minutes 54 seconds after liftoff, when the RL10B-2 engine restarted for a planned 162 second burn and, 3.4 seconds later, suddenly shut down.  The stage began tumbling, but gradually restored its attitude.  Orion 3 separated into a 138 by 153 km orbit, well short of the planned 185 by 25,956 km orbit.  The satellite could not be recovered to a useful orbit and so was declared a total loss.

Another failure investigation began.  Five months after the failure, the investigative board determined that the RL10B-2 engine's combustion chamber had burst during the restart due to defective brazing of a welded reinforcing strip. Pratt & Whitney had to modify its brazing process and its inspection methods.

In many ways, the second Delta 3 failure was a  bigger setback than the first.  According to a November 18, 1999 report by the Associated Press, the two Delta 3 failures cost Boeing more than $100 million.  The failures also delayed the program by more than 2 years compared to original plans.   During the investigation, all RL10 flights, including Atlas Centaur launches, had to be stopped, which resulted in the loss of one payload to Ariane 4.  About the same time, Delta 3 lost its next payload when ICO delayed bankruptcy.  Other potential payloads vanished as economies around the world slid into recession. 

A little-noted 1999 decision would also have implications for Delta 3.  On August 31, 1999, Boeing announced that it would develop Delta 2 Heavy by adding Delta 3's GEM-46 strap on solids to the existing Delta 2 design.   The more powerful motors increased GTO payload by 10% over the standard Delta 2 design.  NASA planned to use the first new Delta 2 Heavy to launch its Space Infrared Telescope Facility (SIRTF). 

Delta 280

Delta 280 First Stage Stacking

A collapsing commercial satellite market forced Boeing to fly the third Delta 3 with a dummy payload named DM-F3 (for Delta Mission Flight 3) on August 23, 2000. The main intent of the flight was to qualify the second stage, which was similar to the yet-to-fly Delta 4 second stage.

This time Delta 3 flew as Delta 280.  Devoid of the usual mission decals, the rocket's large white interstage and payload fairing, separated by the orange band of insulation on the second stage liquid hydrogen tank, provided an unusually simplified appearance.  Only the Boeing logo and a simplified Delta logo on the first stage fuel tank cluttered the image.  The rocket's starkness implied an urgent mission.

The DM-F3 payload was a spool-shaped dummy that weighed 4,348 kg.  The mission was designed to demonstrate a propellant depletion shutdown of the second stage RL10B-2 engine during its second burn.  The press kit listed a targeted 185 x 25,408 km x 27.5 degree subsynchronous transfer orbit.
 

Delta 280 Second Stage/Interstage Stacking

For the first time, Delta 3 lifted off during daylight - shortly after dawn at the Cape.  The rocket repeated its previous performances, shedding its first six GEM-46 solid motors about 80 seconds after liftoff and its second set at about the 160 second mark.  The payload fairing, also a Delta 4 item, jettisoned at about T+227 seconds.  Thirty four seconds later, the first stage RS-27A propulsion system cut off.  Stage separation occurred 296 seconds after liftoff, with the vehicle traveling 4,784 meters/second and climbing past an altitude of 159 km. 

The second stage performed two burns.  The first, 538 second long burn boosted Delta 280 into a 157 x 1,363 km x 29.5 degree parking orbit.  The second burn began 1,315.5 seconds after liftoff, following a 495 second coast to the equator.  The burn was expected to be 163.3 seconds long, but ended a few seconds earlier. 

Delta 280 Payload Fairing with Encapsulated DM-F3 Payload Simulator, Stacking at SLC 17B.

DM-F3 separated into a 180.76 x 20,694 km x 27.5 degree orbit, prompting speculation that Delta 3 had failed again.  But the propellant depletion mission worked differently than the more familiar command shutdown mission.  Atmospheric wind and ambient launch temperature conditions altered the targeted orbit apogee, which was not determined until shortly before liftoff.  Delta 280 ended up with a targeted apogee of 23,404 km.  The achieved orbit was 0.9% below the targeted orbital velocity, within the allowable margin of error (3,000 km for the apogee).  

The variation was unsurprising given the use of a propellant depletion profile on the first complete mission of a new rocket stage.  Propellant management, involving the precise prediction of burn rates for two liquid propellants, is as much art as rocket science.  It typically takes a few flights to tightly "tune" the performance of a new stage fitted with a new rocket engine.  Most new launch vehicles use command cutoff profiles during initial flights.  Such missions have lighter than maximum payloads and are planned to carry a bit of excess propellant.  Delta 280 carried a maximum payload and was fired to push that payload to the maximum possible velocity.
    

Delta 280 Liftoff

Five years after the program had begun, and two years after its first flight attempt, Delta 3 had finally flown true. 

There may have been celebrations at the Cape, but they were likely muted.  The commercial satellite market, along with the world-wide economy, had collapsed.  The next planned Delta 3 launch was more than a year away, at least.   

The End

Delta 3's tough start had cost business, and the collapsing commercial satellite market had caused more to vanish.  But there was another factor in play.  Boeing had poured hundreds of millions of dollars into Sea Launch, an international commercial launch venture, even before Boeing merged with McDonnell Douglas.  Sea Launch Zenit could out-lift Delta 3 by a sizable margin - and its first two launches had succeeded. 

In 2000, Boeing also purchased Hughes Space & Communications, the satellite builder.  Satellite profit margins were sizeable.   Launch services profit margins were slim to none, especially with emerging competition from Russia.  Unlike McDonnell Douglas, Boeing had little incentive to continue to support Delta 3 in a shrinking commercial satellite launch market.

Meanwhile, Boeing's more powerful Delta 4 EELV development program was underway, with first launch expected during 2002.  Delta 4, if it could be brought in at a competitive cost, could make Delta 3 obsolete.  

During May 2001, Boeing reported that it held only five "firm" Delta 3 launch contracts, with none due to fly until 2003.  The company announced that it planned to phase out Delta 3 after no more than about 20 had flown. 

By January 2002, Boeing was reported to be planning to phase out Delta 3 after it had flown out a 9-vehicle inventory.  Two were to fly in 2003 and five in 2004.  The program was to end after 2005.  But by the end of March 2002, Boeing was reported to be considering ending the Delta 3 program altogether.  The company was said to have already cannibalized parts from four Delta 3s. 

After Delta 4 successfully flew for the first time in November 2002, Boeing decided to quietly end Delta 3.  There were no press releases, but in September 2003 Wilbur Trafton, president of Boeing Launch Services, noted in an interview both that Delta 3 was being killed and that Delta 4 was not going to handle commercial business for the foreseeable future. 

Delta 3 RS-27A engines, first stage propulsion sections and LOX tanks if completed, and GEM-46 solid motors were diverted to the Delta 2 and Delta 2 Heavy programs.  Parts from Delta 3 upper stages, including RL10B-2 engines, and payload fairings could be reassigned to Delta 4.  The 4 meter interstages, first stage fuel tanks, and second stage tanks, however, could not find a home.  It is not clear how many such parts were actually manufactured.

Modified SLC 17B continued to host Delta 2 launches.  It also served as the exclusive home of Delta 2 Heavy because only it had been designed to handle the GEM-46 motors.  But by 2010 the end of even Delta 2 was in sight.    

During the late 2000s, Boeing and Pratt & Whitney donated an uncannibalized Delta 3 upper stage fitted with a test-fired RL10B-2 engine to the Discovery Center in Santa Ana, California.  The stage today stands in the corner of a museum parking lot, visible as a sort of billboard, but likely little noted, by passerby on Interstate 5. 

References

"Dynamics Gets Pact to Build 11 Atlas Rockets", San Diego County Business, 5-4-1988.

"Delta Upgrade Challenge to Atlas", Flight International, 2-7-1990.

"McDonnell Douglas Announces Delta 3",   McDonnell Douglas, 5-10-1995.

"Hughes Buy 10 Launches as First Delta 3 Customer", Business Wire, 5-10-1995.

"Russia Pushes for Bigger Market Share, Aviation Week Reports", Business Wire, 1-19-1996.

"Hughes Announces Selection of McDonnell Douglas Delta 3 to Launch Five Satellites for ICO", Hughes, 10-25-1996.

Boeing/McDonnell Douglas Merger Agreement Announcement, 12-15-1996.

"Space Systems Loral Selects McDonnell Douglas Delta 3", Space Systems Loral, 2-13-1997.

"Hot Fire Ignition Test with Densified Liquid Hydrogen Using a RL10B-2 Cryogenic H2/O2 Rocket Engine", Nancy McNelis, Lewis Research Center, AIAA-97-2688, 1997.

"Performance of the Spacecraft Propulsion Research Facility During Altitude Firing Tests of the Delta 3 Upper Stage", Michael Meyer, Lewis Research Center, AIAA-98-4010, July 1998.

"Challenging Pneumatic Requirements for Acoustic Testing of the Cryogenic Second Stage for the new Delta 3 Rocket", Andrew Webb, Goddard Space Flight Center, 9-3-1998.

Delta 259 Failure Investigation Press Releases, Boeing, 8-1998 through 10-1998.

"Boeing Delta 2 Heavy Lift Rocket Developed for NASA Payload", Boeing, 8-31-1999.

"Delta 269 Investigation Report", Boeing, 8-16-2000.

"Delta 280 Mission Book", Boeing, 8-2000.

"Delta 3 Rocket Falls Short, but Still a Success, Boeing Says", Space Flight Now, 8-24-2000.

"Delta 3 Payload Planners Guide", Boeing, 10-2000.

Photos by NASA, McDonnell Douglas, Boeing, Pratt & Whitney Rocketdyne

Return to Thor History Home Page