28 October 2016

George Landwehr von Pragenau's Quest for a Stronger, Safer Space Shuttle

The Space Shuttle Challenger and its booster system moments before they were destroyed. The plume of flame from its malfunctioning SRB is clearly visible. Image credit: NASA
The Space Shuttle Orbiter Challenger was minding its own business on 28 January 1986, working hard to get its seven-member crew and its large satellite payload to low-Earth orbit, when its booster stack betrayed it and everything began to go badly wrong. First, hot gas within its right Solid Rocket Booster (SRB) began to burn through a seal meant to contain it. Soon, a fiery plume gushed from the side of the SRB, robbing it of thrust, and reached out menacingly toward the side of the brown External Tank (ET) and the strut linking the lower end of the SRB to the ET (image at top of post). The plume broke though the ET's foam insulation and aluminum skin, then the strut pulled free of the weakened ET.

Challenger fought back as the ET began to leak liquid hydrogen fuel. It swiveled (the aerospace term is "gimballed") the three Space Shuttle Main Engines (SSMEs) in its tail as it struggled to stay on course. The plume from the SRB, meanwhile, glowed brighter as it began to burn hydrogen leaking from the ET. At the same time, the SRB began to rotate around the single strut left holding it to the ET. That strut was located not far from the Orbiter's gray nose, near the conical top of the errant SRB.

Throughout these events, Challenger's last crew remained oblivious to the technological drama taking place around them. This was just as well, since they had no way to escape what was about to happen to them.

When Challenger at last lost its struggle against its own booster stack, significant events were separated by tenths or hundredths of seconds. Immediately after the right SRB's lower strut came free, the entire Shuttle stack lurched right. Mike Smith, in Challenger's pilot seat, had time for a startled "Uh-oh" less than a second after the lurch. The ET's dome-shaped bottom then fell away, freeing all the hydrogen fuel it contained. The right SRB's pointed nose slammed into and crushed the top of the ET, freeing liquid oxygen oxidizer. The escaped hydrogen blossomed into a fireball that encompassed Orbiter, rapidly disintegrating ET, and SRBs.

Yet the Orbiter Challenger did not explode. Instead, it broke free of what was left of the ET and began a tumble. The aerodynamic pressures the Orbiter experienced as its nose pointed away from its direction of flight were more than sufficient to snap it into several large pieces: the crew cabin, the satellite payload, the wings, the SSME cluster. They emerged from the fireball more or less intact. The SRBs, still firing, flew out of the fireball, tracing random trails across the blue Florida sky until a range safety officer commanded them to self-destruct. The Orbiter's wreckage, meanwhile, plummeted into the Atlantic within sight of the Florida coast.

NASA recovered the bodies of the crew and portions of the wreckage, including the section of the right SRB that had leaked hot gas. The wreckage was turned over to accident investigators.

This 1975 NASA illustration depicts the basic components of the Space Shuttle system. The Orbiter includes three Space Shuttle Main Engines (left) and two Solid Rocket Boosters, one of which is mostly hidden behind the External Tank. The tank comprises (from right to left) a small tank for dense liquid oxygen, a drum-shaped structural support ring/tank separator below the Orbiter's nose, and a large tank for low-density liquid hydrogen.
During a Shuttle launch, the three SSMEs ignited first. This caused the twin SRBs, the bases of which were mounted to the launch pad by explosive bolts, to flex along their entire length away from the SSMEs, then straighten out again just as they ignited. O-ring seals between the cylindrical segments making up the SRBs often became unseated during flexure, then had to reseat to contain hot gases after SRB ignition. Accident investigators concluded that failure of one of those seals doomed Challenger. Even more damning, they found that partial seal failures followed by hot exhaust leaks had occurred on pre-Challenger flights – and had been disregarded by NASA managers.

After Challenger, NASA and its contractors redesigned the SRB joints and seals, added crew pressure suits and a limited crew escape capability, and banned potentially unsafe practices and payloads from Shuttle missions. Yet the U.S. civilian space agency might have gone much farther when it sought to enhance Space Shuttle safety after Challenger.

Even before the accident, NASA had at its disposal redesign proposals that could have made the Shuttle stack stronger and safer. In 1982, for example, George Landwehr von Pragenau, a veteran engineer at NASA's Marshall Space Flight Center, filed a patent application – granted in 1984 – for a Shuttle stack design that would have made the Challenger accident impossible.

Born and educated in Austria, von Pragenau joined the von Braun rocket team in Huntsville, Alabama, in 1957. He became a U.S. citizen in 1963. He specialized in rocket stability and flight effects on rocket behavior. He had, for example, been part of the team that found the cause of the "pogo" oscillations that crippled Apollo 6, the second unmanned Saturn V-launched Apollo test mission (4 April 1968).

In the conventional Shuttle stack, von Pragenau explained, SRB thrust was transmitted through the forward SRB attachment points to a reinforced intertank ring between the ET's top-mounted liquid oxygen tank and its liquid hydrogen tank.  He considered this "indirect routing" of thrust loads to be perilously complex. SSME thrust loads, for their part, passed through the Orbiter to its twin aft ET attachment points on the large fragile hydrogen tank.

By the time he filed his 1982 patent application, von Pragenau had spent almost a decade thinking about how the Shuttle stack might be rearranged to reduce weight and aerodynamic drag, increase stability, simplify paths through which the force of rocket thrust was transmitted, and provide greater structural strength. His 1984 patent was, in fact, not his first aimed at Shuttle improvement.

Von Pragenau's 1974 alternative Shuttle stack. Image credit: U.S. Patent Office

In 1974, von Pragenau had filed a patent – granted the following year – in which he proposed a more slender, more vertically oriented Shuttle stack; that is, one that would mimic conventional rocket designs in which stages are stacked one atop the other. He linked the twin SRBs side by side. Moving the tank for dense liquid oxygen from the ET's nose to its tail placed its concentrated mass nearer the base of the stack, improving in-flight stability. He then mounted the SRBs to the Orbiter's belly and perched the ET atop the SRB/Orbiter combination. SRB and SSME thrust loads were conveyed through struts to meet at the ET's flat, reinforced base.

Von Pragenau's 1982 Shuttle stack design was in some ways a less radical departure from the existing Shuttle design than was his 1974 design. He left the SRBs, ET, and Orbiter in their normal positions relative to each other, but made other significant changes. As in his 1975 patent, he moved the liquid oxygen tank from the ET's nose to its tail and brought the SRBs closer together to improve stability. The liquid oxygen tank became skinny, cylindrical, and almost as long as the Orbiter and SRBs attached to it. The liquid hydrogen tank, fat with low-density fuel, von Pragenau mounted atop the oxygen tank, partially overhanging the Orbiter and SRBs.

Von Pragenau's 1982 Shuttle stack redesign. The numeral "15" points to the rigid thrust structure framework. "34," "35," "36" are Solid Rocket Booster attachment fixtures. These would link to slide rails ("31" and "32") that would run the length of the liquid oxygen tank ("20"). "19" is the liquid hydrogen tank. Image credit: U.S. Patent Office 
Von Pragenau could not tolerate flexing SRBs. He proposed to mount a slide rail on either side of the liquid oxygen tank. Three attachment fixtures on each SRB would link to the slide rails, helping to ensure rigidity. When the SRBs depleted their propellant, pyrotechnic bolts would fire, freeing them to slide backwards down the rails and fall neatly away from the Orbiter/ET stack.

The most important feature of von Pragenau's redesign was a rigid framework – a thrust structure – that would link the bottom of the SRBs just above their rocket nozzles. In addition to holding the SRBs rigidly in place, the thrust structure would transmit SRB thrust loads to the bottom of the ET oxygen tank, which would rest atop the center of the thrust structure. When the spent SRBs slid away from the Orbiter/ET stack, they would take the thrust structure with them.

Side view of Von Pragenau's 1982 Shuttle stack concept. Image credit: U.S. Patent Office
Von Pragenau's concepts apparently exerted little influence on NASA's post-Challenger recovery effort. A likely explanation is that neither of his proposals – if they were known to decision-makers at all – was deemed affordable. In addition to extensive changes in manufacturing tooling, both proposals would have required modifications to the Vehicle Assembly Building, the twin Complex 39 Shuttle pads at Kennedy Space Center (KSC), and even the barge that delivers ETs to KSC. Instead of beefing up the existing Shuttle, NASA studied designs for new shuttles which, for lack of funding, remained firmly in the low-cost realm of CAD drawings, conference papers, and conceptual artwork.

On 1 February 2003, the Space Shuttle claimed another crew. The oldest Orbiter, Columbia, was heavier than her sisters Atlantis, Discovery, and Endeavour, which limited the amount of cargo she could deliver to the International Space Station (ISS). For this reason, NASA largely relegated to Columbia the few remaining non-ISS missions - for example, Hubble Space Telescope servicing.

As they began Earth-atmosphere reentry at 8:44 a.m. Eastern Standard Time after a nearly 16-day life sciences mission, the seven STS-107 astronauts on board Columbia were unaware that, during ascent, a piece of ice-impregnated insulating foam nearly a meter wide had broken free from the ET and impacted their spacecraft's left wing. Ice and foam had broken free from ETs before, but the damage they caused was, after cursory examination, deemed acceptable by Shuttle Program managers. This time, however, the impact opened a hole up to 10 inches wide in the Orbiter's left wing leading edge.

Hot plasma generated during reentry entered the hole and began to destroy Columbia's left wing from the inside out. Observers along the Orbiter's flight path, which cut across the southern tier of U.S. states, reported unusual flashes. Meanwhile, members of the STS-107 crew on Columbia's Flight Deck observed and recorded on video flashes visible outside their windows. In the recovered video, the astronauts appear to realize that the flashes were unusual but show no signs of panic.

Much as Challenger had before it, Columbia fought bravely against the forces destroying it. Onboard computers took account of increased drag on the left side of the Orbiter and sought to compensate to keep it on the flight path. At 8:59 a.m. Eastern Standard Time, however, Columbia tumbled and disintegrated over northeast Texas.

Both of von Pragenau's design concepts placed all or part of the ET above the Orbiter, so one might argue that they would not have prevented a failure resembling that which killed the STS-107 crew. On the other hand, one can be forgiven for speculating that a U.S. civilian space agency provided with the means after Challenger to rebuild the Shuttle system to make it safer might also have evolved an organizational culture more prone to investigating and less prone to tolerating recurring flight anomalies.

Von Pragenau retired from NASA in 1991 after more than 30 years of service. He remained involved in engineering efforts at NASA Marshall Space Flight Center during his retirement. He died two years after the Space Shuttle's final flight (STS-135, 8-24 July 2011), on 11 July 2013, at the age of 86.


Patent No. 4,452,412, Space Shuttle with Rail System and Aft Thrust Structure Securing Solid Rocket Boosters to External Tank, George L. von Pragenau, NASA Marshall Space Flight Center, 15 September 1982 (filed), 5 June 1984 (granted)

Patent No. 3,866,863, Space Vehicle, George L. von Pragenau, NASA Marshall Space Flight Center, 21 March 1974 (filed), 18 February 1975 (granted)

Hampton Cove Funeral Home Obituaries: George Landwehr von Pragenau - http://www.hamptoncovefuneralhome.com/fh/obituaries/obituary.cfm?o_id=2150841&fh_id=13813 (accessed 27 October 2016)

NASA History: Columbia Accident Investigation Board - http://history.nasa.gov/columbia/CAIB.html (accessed 29 October 2016)

NASA History: Challenger STS 51-L Accident - https://www.hq.nasa.gov/office/pao/History/sts51l.html (accessed 29 October 2016)

More Information

What Shuttle Should Have Been: NASA's October 1977 Space Shuttle Flight Manifest

Where to Launch and Land the Space Shuttle? (1971-1972)

What if a Space Shuttle Orbiter Had to Ditch? (1975)

What If Galileo Had Fallen to Earth? (1988)

Dreaming a Different Apollo, Part Two


  1. It does, though of course the various components would play different roles. For example, the hammer-headed LH2 tank in von Pragenau's 1982 patent design was the shroud holding large-diameter payloads and upper stages in the Titan IV. I suspect you know that. But the resemblance is striking.


  2. Well written, bittersweet but fascinating article. Thank you for this.

  3. You are welcome. I cannot help but wonder how many lives and how much money we might have saved if we had made adequate investments in Shuttle improvements over the years. If we meant to keep the Shuttle flying well into the 21st century, we should have sought to mitigate known frailties in the design. The entire booster stack was makeshift - a solution to the Nixon Administration's ardent desire to not fund the Shuttle properly.

    One of my readers posted on my Twitter feed NASA plans for Shuttle life-extension improvements. These included Liquid Rocket Boosters (LRBs) that could have been turned off and which would have made possible a broader range of credible abort modes; a fully reusable unpiloted liquid-propellant booster; new, less labor-intensive, more durable thermal protection systems; auxiliary cargo volumes; Shuttle-C for cargo-only and ISS-assembly flights; etc., etc.

    So, it was a conscious decision, or set of decisions, to leave crews vulnerable throughout the Shuttle Program, and even after two accidents that killed a total of 14 people. I think that's indefensible. I blame NASA to some degree, but the truth is that it followed a path others dictated. Human spaceflight was not accorded the respect and attention it deserves. We remain in this mode of thinking - not adequately funding a Shuttle replacement is the most obvious example.


  4. I'm not sure I can add much more to your comment above. You summed up my feelings quite succinctly about how we could have invested wisely early on in the shuttle program to make it robust over the long-term. I don't believe there will be a viable solution to the challenges NASA faces budget-wise as long as funding is a political football. I'd love to see NASA separated from the yearly budget battles and have their funding allocated in 4-year blocs. It could be timed to reflect the wishes of the American people after each Presidential election. If NASA had this kind of stability with their year-to-year budget, I am confident they could accomplish any task given to them by the government. I know this is mere fantasy on my part at this time but who knows what the future may hold. Thanks for the poignant and informative article on the loss of Challenger and Columbia.

    1. Robert:

      I think we need to start at the start - we aren't going to get preferential treatment for NASA until we make people more aware of what it does and what it could do. Hence the need for blogs like this (ahem) and all the STEM efforts afoot across the country.

      Critical in building awareness is presentation of real information and realistic plans. So much of what we see promoted as realistic space goals is a distraction from real spaceflight. Near-term Mars settlements, to cite perhaps the most obvious example. These "faux goals" are basically indistinguishable from science fiction stories, which I suspect explains a large part of their appeal - but also explains why most people don't take them seriously.

      This is the kind of thing that make me unpopular, sometimes. What I know is that what we have tried since the 1960s hasn't worked to build the space program we want. How, then, can it hurt to embrace a more realistic approach?

      I say we declare for the development of cislunar space. Portray it as a part of Earth. It is just that, after all.

      People worry about asteroids and want to deflect them - I say, "Bring 'em on." If we built up a proper cislunar infrastructure, any asteroid that wandered too close would surely regret it, because Cislunar Humankind would simply devour it - turn it into new facilities, propellants, etc. Basically, a cislunar reach changes our relationship with space in a way that makes the "fsux goals" attainable (if we should choose to pursue them).

      This is hard to make work, not least because folks find it hard to let go of their unrealistic dreams. But related to that, we can't completely give up our unrealistic dreams. Because if we take the proper steps, with the proper expectations, in the proper order, they can become realistic.

      Declaring for a Mars settlement in the next few years is silly. Cislunar development, on the other hand, is already begun. Let's make a list of all the nifty things we can do in the space bounded by the most distant Earth-moon L point. I am sure if folks are made aware of what we could do, they would become inspired. Not least because, as I say, we have been developing cislunar space since the 1960s. It's real, even if nascent.


  5. David, I'm interested in what you have to say about the escape systems that considered and rejected in the early phases of designing the shuttle. I still shake my head at the fact that 14 people had to die get it into our collective heads that space capsule is and always was the safest approach to manned space-flight.

    1. JK:

      This is a complex issue, or set of issues. Shuttle began as a reusable piloted transport meant to be augmented by large expendable rockets. Most everyone assumed that Saturn rockets would do the heavy lifting. As we moved away from that, we compromised safety and added new risk factors.

      So, a big Shuttle Orbiter - part of a fully reusable system - packed full of cargo and propellants had to be less safe than a small crew carrier that could be plucked free of its booster using a proven tractor system. At least the fully reusable system could turn off engines and retreat to a Saturn V-launched space station safe haven in the event of trouble. Some had cockpits that could blast free of a faltering fully reusable Shuttle. Some had jet engines so they were not massive gliders. These approaches reduced cargo capacity and added complexity, however, so were never very popular. It was always considered safer to keep the crew in the Orbiter if possible, and finding places for the jet engines was always tricky.

      Removing second-stage propellants from the Orbiter made it less fragile - but then we mounted it on a makeshift booster system, so we introduced more risk. It was a booster system failure that killed COLUMBIA and CHALLNGER and 14 astronauts, as you know. I hate it when people say "COLUMBIA accident," for example. COLUMBIA performed amazingly given the cards it was dealt. CHALLENGER didn't have time to show its mettle, but it did its best.

      Shuttle II designs generally reverted to the big reusable Booster/reusable Orbiter plan. If I had been able to wave a wand and change everything, though, I'd have developed heavy lift and made the Shuttle a pure crew carrier. Something like the Hermes system. Revert all the way back to the mid-1960s ideas, in other words.

      Is a capsule safer? I think in general it can be more robust, but we have to be careful. Is Dragon safer than Shuttle? I'm not sure. It has a new, barely proven booster and a "push" escape system in place of a tractor system. I much prefer the more obviously Apollo-derived capsule shapes launched on proven rockets. They're simpler and we've an experience base to build on. Of course, we'd also need a stable of rockets and spacecraft - in the latter category, a space tug that could serve as a capsule service module, perhaps - that would enable us to launch cargo all through cislunar space.


    2. I liked Energiya Buran. Fly heads up to ease foam strikes, with no oxygen "ramp: nearby--and LFBs.

      I still do like side-mount, in that I don't trust top mount spaceplanes--pitch loads, bending moments. Side mount allows wider wingspan--and aerobrake disks.

    3. Anon:

      I liked Buran's automated mode and Energia's "built-in" Shuttle-C capability. My impression is that it was not as fragile as our Shuttle, though since we only saw two flights, one with Buran and one without, who knows what failures modes were lurking in the design?


  6. Love your writing as always, David - I have a somewhat tangential question. Soyuz orbital modules have the capacity to be left behind as an additional station module, why do you think this isn't commonly done? Is it a maintenance issue?

    1. P.S. did you ever start a Patreon? I'd be glad to donate.

    2. EE:

      I had forgotten about that. Yes, the Soyuz Orbital Module was left behind at least once, I think on Salyut 7 near the end of its career. It had to be discarded before anything could dock at the port where it was docked, so it couldn't become a permanent station module. I wondered for a time whether leaving it behind was a test of a contingency mode - what if the station docking unit jammed somehow? Separating from the OM would let the crew come home, but would also affect Soyuz reentry operations. If nothing else, the spacecraft would "handle" differently as it moved away from the station and performed its de-orbit burn. I have no proof that that was the real reason for leaving the OM docked to the station, though.

      I need to prove to myself that I can deliver new good posts on a regular basis before I take money from anyone. Thanks for your offer of support - I hope to be able to take you up on that in 2017.


    3. Some mistakes as fas as Soyuz goes. The orbital Modeule cant be left behind as part of a spaxcestation as it doesnt have an second arilock s if it is left in a station it will be useless. The Chinese copy of the Soyuz on the other hand (Shenzu) has this cappability and the Chinese used to left the orbital module in space for testing various technologies. As for the Salyut 7 the thing that was attached (Kosmos something cant remember the number) wasnt a Soyuz but TKS derived capsule. TKS was a planned capsule for supplying the Almaz space station that was designed by Chelomei.

    4. I recommend you look up my 1995 NASA-published book MIR HARDWARE HERITAGE, which I wrote while working for NASA. You can find the PDF online on a couple of NASA sites. It will fill you in on the Soyuz Orbital Modules and large modules attached to Salyut 7, as well as many other details of the Soviet/Russian space programs through 1994.


  7. Where there any proposed fixes that would not have required a major overhaul of the launch pads and other support facilities? The closest thing I can think of was the Shuttle-C. It's a fix in at your not flying people in it and had added value in being able to haul a load more into LEO and your not flying it back. It always seemed to me that it was one of the best lost-opportunities of the shuttle program. I remember that after the Columbia was destroyed the idea was floated to shuttles unmanned to complete it's part in building the ISS. Your point about cause of the tragedies never really being the Shuttle itself is well-taken. It was truly an amazing spacecraft. Do you have any of the illustrations of what the Shuttle II may have looked like?

    1. JK:

      I plan a new Shuttle II post soon. For now, if you google "Beyond Apollo Shuttle II" you should find a post I did on my WIRED blog. There are pictures there of a model of JSC's Shuttle II concept. I have since learned more about the Langley Shuttle II, the design of which is better known and was favored over the JSC Shuttle II. The JSC Shuttle II is sexier, in my opinion, but I plan to give the LaRC design more attention in the new post.

      Regarding the first part of your question - Shuttle-C received an enormous amount of attention over the years. Option C for the redesigned station was a single-launch station launched on a Shuttle stack in place of an Orbiter. It would have been enormous inside. I think Shuttle-C had some "hidden costs" - for example, if you launched a station module, you'd have to include a tug to maneuver it to the station. That tug would have to be funded. Tug was a political football from the get-go, I think because it gave NASA a foot in the door for lots of exciting missions, such as return to the moon and humans in GEO.

      Shuttle-C was a NASA Marshall project, and JSC never liked those. Plus, it would demonstrate that the Shuttle could deliver cargoes without crews, and that's not a message anyone involved in piloted spaceflight wanted to send. And, like the tug, it opened the door to exciting projects. Congress was set on limiting human spaceflight, not on giving it new opportunities, and not even President Reagan had the motivation to push hard to expand piloted spaceflight. Announcing an $8-billion lab Station at the beginning of the 1984 election year was enough for him. Expanded Shuttle capabilities might have come about as part of President Bush I's SEI - he was sincerely interested in expanding piloted spaceflight - but Congress would have none of it. So, Shuttle-C just never got any traction.

      Von Pragenau's fixes were really radical. As far as I'm aware, things like LRBs would have needed facilities mods, but not many pad mods. Again, LRBs were a Marshall thing, which for JSC was enough reason to mostly oppose them.


    2. Thank you. I found it it looks really cool. Funny, how some things never change. NASA is billing the Orion as the Space Craft that will take us to Mars. It won't do much more than orbit the moon on its own but then the general public I'd pretty clueless about what's happening in space. I'm remembering how the Shuttle was hyped-up showing it building giant solar power stations to beam electricity to earth as well a massive space habitats. All of it giving the impression that we would be flying tons of shuttles with little or no turnaround time. Have to love that Sci-fi marketing.

    3. JK:

      Assuming that humans go to Mars anytime soon, Orion (or a derivative) would serve as the ferry for transporting the flight crew to the Earth-Mars transfer spacecraft and as the Earth-atmosphere reentry vehicle at end-of-mission. Depending on the mission profile selected, it might have other uses, too. Will we launch humans to Mars anytime soon? Perhaps to Mars orbit, where astronauts could teleoperate centauroid rovers on the surface and receive sample canisters launched from the surface. I think that's the most plausible profile because it avoids the cost of descent and ascent vehicles and surface assets, ties together the human and robotic sides of NASA, and postpones the day when we thoroughly contaminate the martian environment until we've had adequate opportunities to seek out life and decide whether contaminating the planet is really we want to do. One could reach Mars orbit and return to Earth in something only a little different from Skylab launched on an SLS derivative. I look at the experience we can gain in cislunar space, the hardware we might develop, and the ethical environment likely to develop over the next few decades, and that's what I come up with as the most advanced piloted mission we are at all likely to see in the next 30 years.

      Some NASA art showed the Shuttle doing big assembly jobs in space - when the Dept of Energy got involved in NASA's space solar power plans, it wanted solar sats far bigger than anything NASA had envisioned, which spawned plans for new, larger launchers. Shuttle became seen as the early test-bed for space solar power systems and space assembly at that point. When Reagan took office in 1981, he scrapped all those plans right away.



I like hearing from my readers. No rules except the obvious ones - please keep it civil and on topic.