Historic scrubs and aborts that could happen to Starship

by Sawyer Rosenstein

There are two words that most people never want to hear at a rocket launch, “abort” and “scrub.” As Starship’s orbital test flight gets closer, it may take multiple attempts to get Booster 7 and Ship 24 off the ground. If so, it wouldn’t be the first super-heavy launch vehicle to experience issues before and/or during flight. Here’s a look at some of the aborts across history, some of which could happen during a Starship launch attempt.

Saturn V

The first flight of the uncrewed Saturn V and Apollo capsule was originally scheduled to lift off in early 1967, two years behind schedule. However, it wouldn’t get off the ground until November of that year. On Jan. 27, 1967, a fire occurred during an integrated test at LC-34 with the crew onboard the Apollo capsule sitting atop a Saturn I-B. The vehicle was not fueled at the time and was running on internal power.

During the test, an ignition source, likely from “vulnerable wiring carrying spacecraft power,” sparked in the 100 percent pure oxygen environment inside the capsule. Virgil “Gus” Grissom, Edward “Ed” White, and Roger Chaffee lost their lives when they were unable to open the hatch.

The first Saturn V lifts off on the Apollo 4 mission. (Credit: NASA)

Crews working on the first Saturn V immediately began making changes following the accident, causing further delays. During an inspection of the Apollo 4 vehicle, a reported 1,407 errors were found in the spacecraft. As for the launch vehicle, more than 1,200 problems were found, including finding an extra, out-of-place bolt in one of the J-2 engines.

During its first countdown demonstration test in October, a flight computer failed at L-45 minutes. The test was tried again five days later, only for more computer problems to arise. They would continue working on the vehicle until it finally launched on November 9, 1967, more than 11 months late.

As for in-flight problems, the Saturn V was equipped with a Launch Escape System (LES). If an abort was required, a command could be issued from the ground by one of three positions in Mission Control. Only the Booster engineer, the Flight Dynamics Officer, and the Flight Director could issue the verbal command, which would allow the crew to toggle an abort switch to the “A” side, then the “B” side, return it to neutral, and pull an abort lever.

A test of the Apollo Launch Escape System in White Sands, New Mexico. (Credit: NASA)

At certain points in the flight, the Saturn V did include an automated Emergency Detection System. It would activate within the first few minutes if a structural failure was detected between the command module and instrument unit on the S-IVB stage, if two or more S-1C engines dropped below 90 percent, or if the pitch, yaw, or roll redline rates were exceeded.

The closest the rocket ever came to an abort was on Apollo 12. 35 seconds after launching on Nov. 14, 1969, the vehicle was struck by lightning, creating a static discharge that knocked all three fuel cells, which meant the entire craft was running on battery power. It was struck a second time 52 seconds into the flight, knocking out the attitude indicator onboard and sending scrambled data to flight controllers. However, the rocket remained on course.

The launch was saved by the now infamous phrase, “set SCE to AUX.” Uttered by the Electrical, Environmental, and Consumable Manager (EECOM) John Aaron, this allowed the raw data from the instruments to be converted into signals that could be read in Mission Control. Seeing no serious problems, the mission proceeded as planned.


During a static fire test, Starship broke the record for the most engines ever fired aboard a spacecraft with 31 of its 33 Raptor engines. The rocket narrowly beat the Soviet N1, which had its own share of scrubs, aborts, and failures. Powered by a first stage consisting of 30 NK-15 engines, the N1 was developed to send cosmonauts to the moon and back during the height of the “space race.”

The original plan was to have a vehicle capable of carrying 40 to 50 metric tons of payload to orbit by 1963. Due to numerous production delays, the first test flight was scheduled for Sept. 1968. While testing the rocket on the ground, the first stage oxidizer tanks began to crack, causing them to end the test and roll the vehicle back.

A Soviet N1 rocket lifts off before failing later in the flight. (Credit: RSC Energia)

After a one-day delay due to weather, the first N1 lifted off on Feb. 21, 1969. While the initial liftoff was successful, trouble began after metallic particles became lodged in the gas generator turbine of one of the 30 engines. This led to large oscillations, which eventually caused some engine components to tear off their mounts and created a propellant leak that soon engulfed the tail compartment.

While onboard systems detected the fire, it gave the incorrect signal to stop all engines just under 70 seconds into flight. The Range Safety Officer issued a destruct command shortly after.

Their next test failed less than a second into the flight when an oxidizer pump ingested a slag fragment, exploding an engine and resulting in another fire. All engines were automatically shut down, causing it to crash and destroy the launch pad.

Engine redesigns began while work was underway to repair the pad. Three years later, the vehicle was six seconds away from first stage burnout when a programmed shutdown of six engines led to an oxidizer pump exploding. The Range Safety Officer issued the command to destroy the vehicle.


The Soviets’ next attempt at a super heavy lift vehicle would be known as Energia. Work began immediately after the decision to cancel the N1 rocket in 1976.

The Energia used four strap-on boosters each powered by a four-chamber RD-170 engine burning kerosene and liquid oxygen (LOX), and a central core stage with four single-chamber RD-0120 engines. The main purpose was to launch the USSR’s version of the Space Shuttle known as Buran. Therefore it was designed to carry all payloads mounted on the side of the stack as opposed to on top.

The Energia launcher carries Buran to orbit. (Credit: RSC Energia)

Its first launch in May 1987, Energia-Polyus, would deliver a satellite into an orbit where it could then use its own engines to burn into its final orbit as an upper stage had yet to be implemented. However, due to a computer software error in the attitude control system, the Polyus satellite could not burn itself into its planned orbit and instead reentered over the Pacific Ocean.

Energia did fly successfully on its second flight, carrying Buran into orbit in Nov. 1988. As the Buran program ended and as the Soviet Union began to collapse, Energia was discontinued. However, the boosters from it were then used in Zenit rockets, and more famously, the four-chamber RD-170 engines would be developed into a two-chamber version known as the RD-180, which is still in use today aboard the Atlas V.

Space Shuttle

While the Space Transportation System, or as it’s better known the Space Shuttle, isn’t necessarily a super heavy lifter, its abort scenarios are still significant and comparable to what could be seen during Starship flights.

Over the course of 135 missions, there were multiple delays and scrubs of shuttle launches. In an interview with the University of Central Florida, Phil Metzger, an engineer on the Space Shuttle for part of his 29 years at NASA, estimated that 50% of Space Shuttle launches were canceled.

According to analysis by the Associated Press back in 2007, four years before the shuttle retired, only 47 of the 118 flights so far lifted off on time. More than half of the delays were caused by technical malfunctions, while foul weather made up about a third of the delays.

The record for the most launch scrubs is tied between STS-73 in 1995 and STS-61C in 1986, both aboard Columbia. Those missions each launched on their seventh attempt to get off the ground.

As for aborts, the shuttle had a few scares. It’s important to note that once the solid rocket boosters on the shuttle stack ignited, they could not be turned off and were the last parts to ignite before liftoff.

Space Shuttle Discovery moments after an on-pad abort during STS-41D. (Credit: NASA)

During the maiden flight of Discovery, the shuttle program encountered its first Redundant Set Launch Sequencer (RSLS) abort, which involved an automatic shutdown of the engines after ignition.

A stuck valve prevented LOX from flowing into the combustion chamber of one of the three main engines triggering the abort. The pad deluge system was used to put out a hydrogen fire on the launch pad caused by the free hydrogen that had collected around the nozzles following shutdown.

STS-41D lifted off on Aug. 30, 1984. The lessons learned from this abort would be implemented in future shuttle missions, including improved procedures to “safe” the vehicle.

STS-51F had experienced a similar RSLS in early July 1985. Three weeks later, Challenger took to the skies on a mission that would mark the only in-flight abort of the Space Shuttle program. Just under six minutes into the mission, temperature readings in engine one’s high pressure turbopump indicated it was above the maximum redline numbers, causing it to shut down. With three minutes left in a normal ascent profile, the orbiter was too high and too fast to perform a Return to Launch Site abort, which would have the orbiter land back at the Kennedy Space Center.

Space Shuttle Challenger shortly after lifting off on STS-51F prior to an Abort to Orbit call. (Credit: NASA)

That left two other abort options on the table. A transatlantic abort would see the shuttle land at an alternate landing site, of which multiple were located around Europe. The third option, which Challenger used, was called Abort to Orbit. That meant firing the OMS engines for 106 seconds, consuming a large quantity of much-needed propellant, but permitting the shuttle to continue into a lower-than-planned orbit. Another option, called Abort Once Around, which would have meant completing one orbit before reentering, was never used.

STS-55 saw Columbia perform an RSLS abort in March 1993 after a LOX pre-burner check valve leaked in the final seconds, causing the purge system to over-pressurize. This was a result of a tiny rubber fragment trapped in the propellant valve.

Later that year, Discovery’s number two engine posted a “major component failure,” leading to engine shutdown about one second after ignition on STS-51.

The closest any shuttle came to launching before it aborted was STS-63. On Aug. 18, 1994, the discharge temperature of one of Endeavour’s high pressure oxidizer pumps exceeded its redline value, triggering the abort at T-1.9 seconds. It was so close to ignition that the onboard computers had already switched into ascent configuration, known as 102.

Space Shuttle Endeavour moments before the onboard computer would abort the launch attempt of STS-68. (Credit: NASA)

Improvements to engines meant no other orbiters aborted after engine ignition, although Columbia’s troublesome STS-93 came close, with an abort called at T-7 seconds.

This also makes Atlantis the only orbiter to never experience an RSLS abort.


Much of the hardware and technology created for the shuttle program is now incorporated into the Space Launch System (SLS).

Like the Space Shuttle, SLS uses liquid hydrogen and LOX to fuel the vehicle in addition to two five-segment solid rocket boosters. The vehicle first rolled out to the pad in March 2022. During its first Wet Dress Rehearsal (WDR), a mobile launcher pressurization problem forced a scrub. A second attempt was scrubbed the following day after issues with LOX temperatures and a vent valve became stuck.

While preparing for a third WDR, a helium check valve was discovered to be in a semi-open position because of a small piece of rubber. That third test, which did not include fueling the second stage, ended early after a hydrogen leak was discovered on the tail service mast umbilical plate, a part of the mobile launcher that is located near the base of the rocket.

The liquid hydrogen fueling and data connection lines. This is one of two Tail Service Mast connections to the rocket. (Credit: Jack Beyer for NSF)

After rolling back to the Vehicle Assembly Building (VAB), the fourth WDR indicated another hydrogen leak in the quick disconnect portion of the tail service mast umbilical.

The first launch attempt was made on Aug. 29. Fueling was delayed due to weather, but all proceeded as planned. However, a hydrogen leak was detected, and the amount leaking out was above redline levels. Eventually, the leak was contained, continuing the count until a sensor determined that one of the four RS-25 engines did not reach the right temperature during engine chill. It later turned out to be a sensor issue.

A second attempt was made on Sept. 3. However, a hydrogen leak in the tail service mast umbilical quick disconnect once again forced a scrub. The vehicle would eventually roll back with more delays caused by Tropical Storm Ian. The vehicle finally launched successfully in the early morning hours of Nov. 16.

SLS lifts off on its first flight. (Credit: Julia Bergeron for NSF)


As SpaceX teams prepare for a launch of the full Starship stack, a few on-pad aborts have already been seen during suborbital tests of some of the ships. SN8 would shut down its three raptor engines only two seconds before liftoff on its first attempt on Dec. 8, 2020.

SN10 had a similar problem. The first attempt on March 3, 2021, was aborted after onboard computers detected the engines were producing more thrust than expected.

(SN8 on Pad A awaiting launch. Credit: SpaceX)

The full stack of ship and booster has performed a wet dress rehearsal and a static fire test, where almost all of the 33 raptor engines ignited for a few seconds. No orbital launch attempt has been made as of yet, but Elon Musk has said the plan is to attempt that in April 2023.

Starship does not have any abort modes that have been made publicly available. If something were to go wrong during a flight, the vehicle is equipped with an autonomous flight termination system, meaning if specific parameters reach a critical threshold, the vehicle will destroy itself. Musk has stated in the past that he hopes for the test to clear the launch pad.

(Lead image: Booster 7 fires 31 Raptor engines. Credit: Nic Ansuini for NSF)

Related Articles