Safety first: Ten years since STS-114’s return to flight misssion

by Chris Gebhardt

Ten years ago today, the Space Shuttle Discovery left the Kennedy Space Center on STS-114. The flight, filled with hopes, aspirations, and emotions, marked the first step in returning the Shuttle to flight following the loss of Columbia and marked the beginning of a new journey and vision for NASA. But getting to launch day for STS-114 was a difficult and safety-conscious endeavour.

STS-114: From Atlantis to Discovery

As NASA moved to begin implementing the safety recommendations mandated by the Columbia Accident Investigation Board (CAIB), progress began to take hold.

As safety modifications coalesced and mission planning successfully split the Return To Flight (RTF) mission into two flights (a redesigned STS-114 and a newly added STS-121), NASA began to consider exactly when the Shuttle fleet could return to flight.

Based on the forward progress and the need to maintain safety above all else, NASA selected 13 May 2005 as the realistic target launch date for the STS-114 mission on Atlantis.

With that date in place, preparations proceeded around the country, with the major milestone of Solid Rocket Booster (SRB) stacking commencing at the Kennedy Space Center in the Vehicle Assembly Building (VAB) for STS-114 in December 2004.

At the Michoud Assembly Facility (MAF) in Louisiana, engineers and External Tank officials diligently removed all of the identified foam hazards from External Tank-120 (ET-120), the tank selected to fly the STS-114 mission.

But this removal caused a debate, specifically toward the location on the Tank where foam had not been removed.

2015-07-26-021228The CAIB specifically mandated the removal of foam from the bipod strut attach area.

However, some within NASA questioned whether the Protuberance Air Load (PAL) ramp foam strip running down the pressurization lines on the left-hand side of the tank posed a significant hazard to the wing leading edges of the Shuttle orbiters.

In the end, NASA decided to proceed in a phased approach to foam removal.

The rationale was not to remove too much at the same time until a better understanding of how the foam reacted during flight could be gained.

2015-07-26-015435To further this understanding, NASA placed a camera on the External Tank to look down over the left-hand side of the tank where foam shedding events could be observed, documented, and analyzed.

Thus, ET-120 and its subsequent tank, ET-121, did not have their PAL ramp foam removed.

ET-120 was subsequently declared complete at MAF and shipped to the Kennedy Space Center for mating to the SRB stack for use by Atlantis during launch.

But in the end, it was not to be for Atlantis.

As the Orbiter Discovery went through Orbiter Maintenance Modification (OMM) at the Kennedy Space Center, an improperly installed gear within her Rudder Speed Brake (RSB) on her vertical stabilizer was discovered.

2015-07-26-021458This prompted concern that the same gear might also be improperly installed on Atlantis and Endeavour.

Subsequent inspections of Atlantis found that her RSB gear was improperly installed as well.

And now NASA faced a dilemma. Discovery’s RSB gear had been corrected during her OMM. Thus, Discovery did not need any repairs. Atlantis did.

Unfortunately for Atlantis, the amount of time necessary to affect the repairs would have meant missing the May launch window for STS-114.

This caused NASA to consider swapping Atlantis and Discovery, placing Discovery on STS-114 and Atlantis on STS-121.

However, simply swapping Atlantis and Discovery was not a simple decision.

2015-07-26-021551The Shuttle construction flight sequence to Station had been carefully choreographed to place each Shuttle Orbiter on a mission it could actually fly from a weight standpoint.

In short, while Discovery, Atlantis, and Endeavour were all similar in weight, Discovery weighed the most and therefore could not be used on the payload weight-heavy flights to deliver the massive solar array trusses (S3/S4 and P3/P4) to the ISS.

One of these solar array and truss flights was STS-115, the flight destined to follow STS-121.

With Endeavour undergoing a mandatory orbiter modification down period, she was unavailable to fly STS-115. This meant that Atlantis was the only available vehicle for STS-115.

2015-07-26-021713Thus, swapping Atlantis and Discovery for STS-114 and STS-121 would have meant Atlantis flying back-to-back missions on STS-121 and STS-115.

It also would slightly complicate matters as part of NASA’s new safety plan in the post-Columbia era to have a Shuttle ready to launch to rescue a crew stranded by a crippled Orbiter at the International Space Station.

The Orbiter to fly STS-121 would be the rescue vehicle for STS-114. Likewise, the vehicle for STS-115 would be the backup for STS-121.

But if Atlantis flew both STS-121 and STS-115, the launch of STS-121 would have to be delayed by enough days to let Discovery get through enough post-STS-114 ground processing to enable her to be ready to fly a rescue mission for STS-121 if needed.

However, with all these considerations, NASA decided that it would be easier to swap Atlantis and Discovery for the STS-114 and STS-121 flights and simply incur the launch delays with STS-121.

With that decision, Discovery was now the return to flight vehicle for yet a second time in the Shuttle program – as she also flew the 1988 return to flight mission from the Challenger disaster.

Getting to the pad: Additional precautions before launch

With ET-120 mated to the SRB stack in the VAB, technicians prepared for the next major milestone toward launch: Rollover of  Shuttle Orbiter Discovery from her Orbiter Processing Facility to the VAB.

Rollover occurred in the extreme early morning hours of 29 March 2005.

2015-07-26-021806Once inside the VAB, Discovery was connected to the hoisting mating sling/crane and lifted into High Bay 1 where she was lowered next to ET-120 and hard mated to the tank by 30 March.

Meanwhile, in High Bay 3 next door, SRB stacking operations were well underway for the STS-121 RTF-2 mission, a stack that would serve as the rescue vehicle for Discovery’s crew in case Discovery herself was damaged during launch or while on orbit.

Over the next seven days, Discovery was connected to her support equipment on the Mobile Launch Platform and completed integrated vehicle system end-to-end testing.

With all integrated tests showing a good and healthy vehicle, Discovery and the STS-114 stack began to slow crawl to Launch Pad 39B on 6 April.

2015-07-26-022111Arriving at the pad in the overnight hours, technicians and processing engineers began preparing Discovery for a tanking test on 14 April.

NASA engineers and mission managers designed the tanking test to help understand the ice build-up environment on the tank during fueling and help gain more information about how ET foam reacts to the contraction of the tank’s diameter under cryogenic conditions during fueling operations and subsequent re-expansion following de-taking.

At the time, very little was known about how the potential stress of this contraction and re-expansion on the foam could affect its bonding ability to the tank’s surface.

2015-07-26-022243Also unknown at the time was how much this cryo contraction and re-expansion process played into the creation of vacuum pockets within the foam that could subsequently expand and pop during ascent (a phenomenon called cryo-popping), causing a foam shedding or liberation event.

However, while gathering information on these issues was the primary goal of the tanking test, it was an unexpected issue during the tanking test that caused the most concern for STS-114 – an issue that would plague the Shuttle program through STS-122 in Dec. 2007/February 2008.

As NASA engineers began pumping 536,000 gallons of liquid hydrogen and liquid oxygen into the ET on 14 April, two of the four low level liquid hydrogen sensors (known as Engine Cut-Off (ECO) sensors) at the bottom of the liquid hydrogen tank registered “dry” when there were thousands of gallons of liquid hydrogen in the tank.

The ECO sensors record liquid hydrogen levels in the tank during ascent and serve as part of the system to verify fuel levels and trigger emergency shut down commands to the Shuttle’s main engines should a fuel depletion event occur prior to a nominal end of ascent climb.

At the time of the STS-114 tanking test, NASA flight rules mandated that all four ECO sensors be functioning properly before the Shuttle could commit to launch.

The fact that two ECO sensors malfunctioned during the tanking test and that a pressurization valve that maintains ET pressurization levels in the final two minutes of the countdown cycled twice its normal amount prompted concern regarding this ET-120.

NASA kickstarted an investigation into both issues as the ice debris evaluation team reviewed the recorded performance and ice buildup on ET-120 during the tanking tests.

2015-07-26-115710During this ice debris buildup review, the teams noted potentially dangerous ice buildup on the bellows along the liquid oxygen feedline that takes liquid oxygen from the top of the ET to the engine compartment of the Orbiter during launch operations.

All of these issues contributed to a rather quick decision by NASA to take the time necessary to understand the issues and therefore miss the May launch window for Discovery and STS-114.

Initially, NASA tracked the ECO sensor issue to a wiring problem and committed to repairs of that system.

Likewise, NASA also believed that the extra cycling of the pressurization valve was due to a new design of the system.

To validate the ECO sensor corrections and the pressurization valve, NASA conducted a second tanking test on Discovery on 20 May 2005.

2015-07-26-120226During the second taking test, the ECO sensors functioned as expected; however, the pressurization valve again cycled twice as many times as usual.

The results of this test and an increasing concern from the ice debris team lead NASA to an unusual decision: Discovery would not launch with this SRB/ET stack.

Instead, Discovery would roll back to the VAB, be de-mated from this SRB/ET stack, and moved to the SRB/ET stack originally scheduled for STS-121.

2015-07-26-120358The move would allow NASA to fly STS-114 on an ET outfitted with heaters on the liquid oxygen bellows to prevent ice buildup in that region.

With the decision made, Discovery rolled back to the VAB on Thursday, 26 May.

Discovery was then de-mated from ET-120 and lowered into the transfer aisle of the VAB on 31 May.

On 7 June, Discovery was picked up by the hoisting mating crane and lifted into High Bay 3 of the VAB where she was attached to her new SRB/ET-121 stack.

With re-stacking complete, Discovery rolled from the VAB on 15 June 2005 for her second trip to her seaside launchpad.

Back to Pad-B: More ECO sensor trouble

With Discovery safely back at the pad, NASA engineers turned to a final set of debris review risk assessments, which classed the debris risk environment from foam and ice at “uncertain but acceptable.”

This classification stemmed from the fact that NASA literally could not implement all of the mandatory safety recommendations from the CAIB because of the way those recommendations were worded.

2015-07-26-122232Specifically, the CAIB had mandated that NASA eliminate all ET debris shedding.

But this was actually impossible to do. What NASA could do instead was take all steps possible to mitigate the potential for foam debris liberation from the tank.

(In fact, every single External Tank launched with the Space Shuttle from STS-1 through STS-135 liberated pieces of foam. However, from STS-121 through the end of the program, those liberation events became increasingly minor, with the last three tanks flown going down as the “cleanest” ever to reach orbit.)

2015-07-26-122344Finally, on 29 June 2005, NASA managers from around the country convened at the Kennedy Space Center for the Flight Readiness Review for the return to flight mission.

After two days of engineering presentations, questions, and thorough reviews, NASA officially declared Discovery ready for liftoff on 13 July 2005 at 3:51 pm.

Discovery’s seven crewmembers flew to the Kennedy Space Center late Saturday 9 July for the start of formal launch operations.

Discovery then officially entered countdown operations at 6 pm EDT on 10 July 2005.

Finally, launch day arrived.

2015-07-26-122421Fueling of Discovery’s ET began a few hours later than planned due to a heater malfunction on the launch pad.

With the heater repaired, fueling proceeded without issue.

Through the morning and early afternoon hours of 13 July, all appeared well with no technical issues that would delay the launch.

Discovery’s flight crew left the Operations and Checkout building on time and arrived at the pad to begin boarding Discovery for the historic trip to space.

But an issue seen months before would not allow launch on this day.

At 1:30 pm, just 2.5 hours prior to the scheduled launch, one of the four ECO sensors at the bottom of the liquid hydrogen tank malfunctioned.

With a repeat of a problem they thought was fixed, NASA managers wasted no time in scrubbing the launch for safety concerns at 1:32 PM.

Troubleshooting the ECO sensors:

Engineers and managers spent the next seven days trying to track down the cause of the ECO sensor No. 2 failure during the 13 July launch campaign.

2015-07-26-122625But the issue remained oddly elusive, with no obvious signs of why the ECO sensor failed found.

Eventually, NASA opted to proceed toward a Tuesday 26 July launch date for Discovery after a plan to exchange ECO sensor No. 2 and No. 4’s  electronics was put in place in the hopes that it would fix the problem – with extensive monitoring during fueling and launch operations.

With the plan in place and the swap of the electronics complete, NASA proceeded toward liftoff of Discovery on 26 July 2005 on the return to flight mission.

STS-114: The Shuttle’s first return to flight from Columbia

In the early morning hours of 26 July, fueling operations of the ET began and proceeded nominally with no ECO sensor anomalies noted.

2015-07-26-122659Discovery’s flight crew left the Operations and Checkout building had arrived at the launch pad on schedule and began climbing into what would be their home for the duration of the mission.

As the countdown proceeded through the 2.5 hours to liftoff mark, the ECO sensors continued to behave.

By 9:45 am, the countdown reached the T -9 minute and holding mark when NASA launch officials and flight managers finished all necessary work before proceeding with the terminal count.

With the final series of checks and polls complete, the final “go” for launch was given by Launch Director Mike Leinbach to Commander Eileen Collins on Discovery.

“Our long wait might be over. On behalf of the many millions of people who believe so deeply in what we do, good luck, Godspeed, and have a little fun up there,” said Leinbach.

Upon hearing that all was ready, Commander Collins responded, “And thanks to you, to the launch team, and to everyone in the Shuttle Program. The crew is go for launch.”

2015-07-26-122847The final nine minutes of Discovery’s countdown commenced at 10:30 am EDT and culminated without issue with the ignition of the Solid Rocket Boosters and a lift off from pad 39B at 10:39:00 am.

As Discovery left her launch pad to return to the Shuttle fleet to active service following Columbia, she did so completely coincidentally at the exact same numerical time 10:39:00 am that the Space Shuttle Columbia had left her launch pad 2.5 years earlier at the beginning of the STS-107 flight.

As Discovery began her triumphant return to space, an unexpected event happened within the first two seconds the flight.

For the first time in the Shuttle Program’s history, Discovery hit a bird on her way off the launch pad.

2015-07-26-122940The large turkey vulture which had been near the pad as Discovery’s engines roared to life, swooped into the flight path of the top of the External Tank, which speared the bird approximately 2.5 seconds after lift off.

The large bird, which could have caused damage to Discovery’s delicate thermal protection system, thankfully fell along the backside of the External Tank and did no damage to the flight vehicle.

Nevertheless the incident, analyzed through post launch imagery, caused NASA to implement several avian safety precaution tools for the remainder of Shuttle Program.

2015-07-26-124323But for Discovery, the bird was not a problem.

The vehicle rolled gracefully onto course and began the initial climb through the dense lower atmosphere.

As millions of people watched and hundreds of cameras followed Discovery’s climb uphill, all systems were nominal.

Two minutes 1.8 seconds after liftoff, the Solid Rocket Boosters completed their push of Discovery and separated cleanly from the External Tank.

But just a few seconds later, the biggest concern and potential fear for the mission happened.

2015-07-26-124448Two minutes 7.1 seconds after launch, while Discovery was at the upper bounds of the area of atmospheric aerodynamic concern, a significant chunk of foam liberated from the PAL ramp of the External Tank and narrowly missed Discovery’s left wing.

The near-miss was captured in good clarity by the newly installed camera onboard the External Tank.

The liberation event was not seen by the public or NASA during the live broadcast of the launch as the camera angle used at the time was from a long-range tracking camera south of the pad.

But post-launch review of all imagery quickly revealed the event to NASA.

2015-07-26-124620While Discovery, her crew, and all of her new inspection equipment made it safely to orbit, NASA was not happy with the liberation event and the safety hazard it posed.

Demonstrating once again their extreme commitment to safety, NASA ordered a grounding of the Shuttle fleet the very day after Discovery’s launch.

But importantly, for Discovery and her crew, they were safe and on orbit.

It was the first day of a near two-week mission more than four years in the making.

(Images via NASA and L2).

Related Articles