AS I was watching coverage of the last shuttle launch I saw a story about the factory in Utah that made the solid fuel booster. In the back of my mind I recall a story that one reason for the explosion was that the booster was made in sections so it could be transported by train. Thus it needed the O rings between sections. If I remember, there were was another company that was going to make it in one piece and ship it by barge, but political considerations made NASA choose what they considered a less attractive proposal.
I haven’t seen this particular point come up before, but related points are often an issue in discussions of pork spending. Sometimes, you really do have a situation where something could productively be done anywhere, or doesn’t really need doing at all, and a politician will lobby to have it done in their district or state. But most of pork is for things where a particular state really would be the best choice for something, and the politician is just trying to make that known. In this case, the company in Utah was the best choice for making the solid rocket boosters, and so the senator from Utah pushed for them to be chosen. But if they hadn’t been the best choice, he’d have been pushing for some other Utah company that does something completely different, instead.
The story is referring to Frank Moss (D-UT), chair of the U.S. Senate Committee on Aeronautical and Space Sciences from 1973 (Thiokol won the contract in 1974). The NASA Administrator at the time (James Fletcher) had also lived in Utah, but had previously been an executive at Aerojet, a Thiokol competitor located in California. These interconnections and apparent conflicts are typical both then and now. For example, former astronaut Scott Horowitz retired from NASA in 2004, took a senior executive position with ATK (successor to Thiokol) for a few months, then was named the NASA Associate Administrator for the Exploration Systems Mission Directorate, a directorate which then awarded a huge contract to ATK to develop the Ares I.
The only thing remarkable about this is how unremarkable it is. These are never arm’s-length transactions. It is as if you wanted to buy a car, but every dealership in town was owned by someone you personally knew or were related to. There would be a similar story to tell no matter which company had won the contract. If the general public ever takes an interest in what goes into this type of decision, there would be a revolution.
A monolithic motor that big would have required money and time to develop, with no guarantee it would have worked. In contrast, segmented motors had already been operational in the Titan III program for several years.
Widening the circle a little, one might ask why solids were used at all on Shuttle. They are known to be riskier than liquids. The answer to that has to do with budgets - originally the boosters were going to be liquid flybacks, but they would have been new technology and cost too much in up-front development even though they would have been cheaper and safer in the long run. Life is full of compromises.
The segmented Thiokol motor that doomed Challenger worked just fine as long as it wasn’t too cold. Thiokol engineers knew the motor’s limitations and recommended against launching that day, but were overruled by their managers, who were under pressure from NASA to specifically launch in the cold and to generally not delay the launch any longer.
Not really. There are certainly some risks that solid rockets are subject to which liquid-fuel rockets are not, but the same is also true in the other direction. And I’m not sure we’ve actually had enough accidents to tell which set of risks is worse.
Call-out here to Stranger on a Train, in case he does a vanity search.
I have experience as a solid rocket motor engineer and in launch vehicle risk assessment, so I’m pretty familiar with all the arguments pro and con.
Roughly speaking, risk is likelihood times consequences. In terms of likelihood of failure, solids are actually a little better than liquids historically, even though solids can’t be hot-fire acceptance tested. But in terms of consequences, liquids win easily. There are many instances of liquid rocket failures on both manned and unmanned missions that have not resulted in loss of mission or crew. An LRE can shut down in flight and the vehicle can still make orbit, or there can be time for the crew to escape.
With a LRE, you get a chance to see if the engine is running correctly before lifting off, and if it isn’t you can usually shut it down safely. Plus, the propellant tanks are designed to leak before bursting, i.e. fail gradually. With solids, once they’re lit you are either going for a ride or getting blown into hash. Their failure modes are all catastrophic, except possibly for a late-stage nozzle or insulator ejection.
From David Jenkins’ excellent Space Shuttle: The History of the National Space Transportation System The First 100 Missions, 3rd Edition*Originally, many concepts had been based around a pair of pressure-fed liquid boosters, one on each side of the external tank. But by the end of 1971 this configuration had also largely disappeared as estimated development costs had continued to soar. The primary reason for the fall of the pressure-fed booster, MSFC’s future employment not withstanding, would be that nobody had ever developed a large pressure0fed booster before, and that significantly increased program risk. On the other hand, the Air Fore had 120-inch solids in production for the Titan IIIC, and various Air Force and NASA experimental programs had demonstrated 156-inch, and indeed 260-inch, solids. There appeared to be little, if any, development risk for a solid rocket booster…In the end, development cost estimates for the solids were approximately $1,000 million less than for the pressure-fed booster, but both were within the realm OMB had given NASA. However, on 9 February 1972 Caspar Weinburger wrote James Fletcher reminding him that the NASA budget was unlikely to exceed $3,200 million, and that “We [OMB] also fully expect NASA to develop a shuttle system within the $5.5 billion estimate.” If solid boosters were selected, the difference in development costs would create a comfortable management reserve to pay for unexpected problems with the SSMEs or thermal protection system. In retrospect, it was a very wise decision.
There was also a constant exchange of data between NASA and the contractors during the final gasps of the Phase B Double Prime studies. Many of these meetings concerned the possible choice of boosters, and George Low commented that these briefings "yielded the recommendations for each contractor that were most predictable based upon vested interests.
Boeing, still part of the Grumman team, had developed the S-IC [Saturn V Stage 1] and continued to champion it…In contrast, Lockheed was a major manufactured of solid rockets [the Polaris/Posieden/Trident Fleet Ballistic Missiles] and had proposed a variety of sizes and shapes…North American did not appear to have a favorite, somewhat surprising since their liquid rocket engine company] Rocketdyne was thoroughly engage in SSME competition at the time.
McDonnell Douglas liked solids. As a company, it had long experience with liquid rocket engines, having built the Thor and Delta launch vehicles, and the S-IVB stage for Saturn. But the company was also familiar with solids, and used them extensively to augment the thrust of the Delta.*
As with most of the Shuttle procurement, there were certainly political factors in play. It is notable that some components for the Shuttle came from every state in the Union. On the other hand, it is inarguable that the only three realistic competitors for manufacturing of solid rocket motors of this size were Aerojet General (manufacturing in Sacramento, CA), Morton-Thiokol (Promentary, UT), or possibly Hercules Powder Company (Magna, UT). United Technologies Corporation (later a division of Pratt & Whitney before a catastrophic propellant mixing failure resulted in the company being shut down) had built the Titan IIIC SRMs, which were the largest segmented solid rocket boosters used to day, but did not have the capacity to produce the Shuttle SRBs. Aerojet recommended using facilities adjacent to a shipping channel, allowing for sea transportation of a monolithic grain; however, the environment was not ideal for manufacturing and storage of solid propellants, and the difficulty of delivering a booster to a site where it can be qualified (which, for an SRM requires firing of a series of full scale assets at a range of bulk grain temperatures) made the logistics complex. On the other hand, Morton Thiokol (which would later be diversted and merge with nearby Herecules) had the ability to manufacture and test large motors at their Promentory facility. Thiokol also had a willingness to invest a significant amount of their own funds into developing the facilities for manufacturing and processing the Shuttle SRBs, and had the advantage of promising technology transfer between the SR-118 (LGM-118A Peacekeeper Stage I) and an advanced lightweight composite case version of the SRB for Blue Shuttle polar orbit use out of VAFB SLC-6. In hindsight, they were the best choice, albeit not without warts, particularly in their management structure.
The record of who said what to whom seems to be somewhat muddled as different players relate different stories, but SRB Program Director Allan MacDonald’s Truth, Lies, and O-Rings: Inside the Space Shuttle Challenger Disaster lays the blame pretty squarely on NASA MSFC management (and one particular individual who I will not name) which pressured Thiokol management into agreeing to the launch even as several engineers, led by Roger Boisjoly, explicitly warned about a failure of the o-ring field joint seals at cold temperatures. (The launch occurred at temperatures colder than the SRB had been qualified to, and the coldest launch temperature to date.) Boisjoly had previously drafted a memo on this issue and headed a task force to investigate the issue but was essentially ignored by management. Thiokol, which was up for another booster buy to support fifty odd missions, may have felt pressured into agreement, although no specific threats were made. (Readers of the book should bear in mind that MacDonald has a very blatant axe to grind against Thiokol management, which unquestionably punished him for speaking up at the Rogers Committee hearings despite orders to defer questions to corporate legal counsel.)
The Titan SRMs have been previously discussed as experience with large segmented solid boosters. However, there was a more limited amount of experience, with 25 Titan IIIC flights and fewer than 100 test firing of large segmented cases total. Segmented cases have additional failure modes (in addition to field joint blow-by or burn-through) that require unique solutions. As it happens, the 7-1/2 segment SRM on Titan 34D-9, the first Titan launch after the Challenger Disaster, also suffered catastrophic failure shortly after launch due to case burnthrough. Although the design of the field joint was different (one o-ring, with the tang and clevis reversed from the Shuttle SRB) and the failure mode was different, the fact is that assumptions about the operation and safety of segmented solids were not as well validated as people in the field thought. Specifically, the Aerospace Corporation (which oversees all large Air Force space launches) claimed that there were no indications of o-ring blow-by or joint failure in Titan SRMs during a Challenger review prior to the loss of 34D-9, but post-failure investigation indicated a long history of blow-by and near failure with those motors. The SRM went through a reliability redesign and was eventually replaced by the Hercules (later Alliant Technologies/ATK after merger with Thiokol/Cordant) Solid Rocket Motor Upgrade, which reduced the number of segments while increasing propellant load out.
It should also be noted that the cold ambient temperature alone may not been responsible for the failure. The launch of STS-51-L experienced some of the most pronounced wind shear (changes in direction of the wind with altitude) of and Shuttle launch. In addition, CFD analysis indicated that cryogenic exhaust from the liquid oxygen tank may have collected near the starboard-side SRB prior to launch, and specifically in the area of the failed joint. Although partial blow-by of the o-ring seals occurred previously, STS-51-L experienced a perfect concurrence of events that led to the failure. Even with the blow-by, which ruptured the hydrogen tank and cut through a strut that resulted in the External Tank coming loose from the orbiter and vomiting its contents, and the inadvertant release of the SRB into free flight, the right-hand booster performed several end-over-end rotations without breaking up prior to being flight terminated by the Range Safety Officer. The design of the booster was robust; the failure was due to a tiny gas of jet in the particular location where it damaged the ET and attachment strut. As we note regularly during readiness reviews, a rocket launch requires several thousand things to go right to a precision of many significant figures, and in many events only one small thing has to go wrong to end up with in a million little fragments falling back to Earth.
Actually, if you look at the basic statistics of all large rocket vehicles muddled together, solid rockets come out way ahead in terms of reliability (as judged by a loss-of-mission or loss-of-crew event). This is only natural; many heavy lift liquid systems use strap-on solids, so you may get between two and ten solids for one liquid. Plus, many strategic rocket systems that require high reliability for deterrence value have been built, tested, and launched in numbers that dwarf all liquid propellant rocket vehicles. Segmented solids, of which there are only a few design examples, are somewhat less reliable, although it should be noted that the Shuttle SRB (and RSRM) have flown a total of 270 times (plus the SRB-based five segment Ares I-X) with only one incidence of catastrophic failure, which compares favorably in terms of reliability to any other space launch booster in its lift class.
It is true that, unlike liquid engines, the main propulsion system cannot be tested before flight, as the propellent grain is the motor chamber and is consumed during use. Testing before flight is limited to inspection (including x-ray), testing out the thrust vector control system (to a limited extent, as most modern vectorable solids use flexbearings that cannot be fully articulated in an unpressurized state), and doing runthroughs on all of the avionics. Liquid engines can be mounted on a stand and tested before flight; however, liquids are almost never tested in the as-flown stage configuration, and are significantly more complex in their mechanical components (turbopumps, pre-heaters, et cetera). As a result, solid motors for critical applications are built and qualified in production lots, where a sample size is tested as is done with ordnance to assess performance.
In general, solids are good for getting a high level of thrust early in burn where propellent-weight-normalized performance (measured in specific impulse) is not critical and you want to get a heavy vehicle moving quickly to minimize gravity losses; however, they have some drawbacks that are not related to reliability or lack thereof that make them a questionable choice for future man-rated space vehicles. One significant negative of using solid boosters on a man-rated system is the fact that they cannot readily be shut down or thrust terminated, and a vehicle mounted like the Shuttle cannot separate or abort during SRB burn, which makes virtually any booster failure a catastrophic failure; this is particularly true in a configuration like the Shuttle, where you have to have the action time and thrust levels equalized to within a fraction of a percent, lest an imbalance rip the system apart (which is essentially what happened with Challenger). There are also the high thrust variations and lateral vibrations due to self-induced combustion oscillations within the motor grain (generally near the beginning of burn) that require a strong connecting structure and may develop very aggressive environments from which a crew or payload have to be isolated, which is one of the reasons that there were so many difficulties with the Ares I rocket.
Is that sufficient and painfully pedantic detail to answer the question of the o.p.?
Is it possible that the OP is referring to the story that originally the company that made the SRB gave a no-go for launch based on the recommendation of the engineers but that eventually the company decided (for some nefarious reason involving the contract coming up for rebid IIR the legend C) the company eventually gave the go-ahead.
No, I was referring to a barged vs train-hauled solution for the booster, with the latter needing to be in sections while the former could be all one piece. But it looks like I was misinformed and/or mis-remembered.
Not sure I agree with this regarding loss of crew. Historically there is one LOC due to SRM (Challenger) and none due to LRE, at least that we have any insight into. There are, on the other hand, several examples of LRE failures that the crew survived and even some that didn’t affect the mission. Apollo 13 had an early shutdown on S-II that had no impact on the mission.
Of course, the above observations are of very limited statistical significance. Some people have looked at the totality of all propulsion failures (manned and unmanned) and concluded that LREs are riskier to crew; for example, saying that if there had been a crew on AC-70 that may well have been a loss of crew due to LRE. But that requires you to assume that the comparison between SRMs and LREs for unmanned missions hold up for manned missions, which is a leap.
AFAIK all the N-1’s were unmanned tests and their failures did not result in casualties (you may be thinking of the 1960 R4 “Nedelin incident”).
OTOH there was one manned Soyuz staging failure (18-1, 1975) and one manned launch pad explosion (T-10-1, 1983) where the abort and escape procedures had to be activated, no fatalities.
When I saw “Politics” in the thread title, an issue came to mind that hasn’t been mentioned yet.
January 28, 1986 (the date of the failed launch) was the same day Ronald Reagan was scheduled to present his State of the Union Address to a joint session of Congress. There was a huge political incentive to perform the (already delayed) launch on that date. (Even admitting this to be a factor, it may not be clear which Reagan aide, if any, pressured NASA into launching.)
I was making a gross comparison between solid propellent space launch and large ballistic rocket systems and liquid propellant systems without reference to manned or unmanned payload. There has only been one man-rated space launch system that has used solid propellant rocket boosters for ascent (the STS), and there are only three man-rated families of launchers overall that have been flown in sufficient quantity in a common configuration to make any useful degree of statistical reliability estimates (STS, the Titan II-Gemini Launch Vehicle, and the various configurations of Soyuz launchers). However, on an overall basis of all large rocket launch systems, solid propellent systems have been flown in far more quantity in common configurations, and (for fully qualified designs) have rates of predicted reliability based upon flight history that vastly exceed that of liquids. I would agree that liquid propellant systems have more survivable abort modes than solid motors, especially systems that use multiple solids, as described previously, and in general, are a better choice for man-rated systems.
The S-II center engine was shut down due to flight oscillations (a common problem with liquid propellant systems that can’t be characterized in ground testing), but the remaining four engines functioned fine and were operated for a longer duration to obtain the same total impulse. Other than a small adjustment on the S-IVB trans-lunar injection maneuver, there was no change in the mission.
This was investigated extensively by both the Rogers Commission and the media. No indication of any pressure by the White House to conduct the mission on 28 January. Richard Feynman (who was on the Rogers Commission and specifically investigated this claim) goes into extensive detail on this in What Do You Care What Other People Think?: Further Adventures of a Curious Character. NASA was under pressure to maintain their launch manifest, and delaying a launch meant pushing other missions “to the right” (i.e. later), but the real issue was that NASA Marshall management, which made the go-no go day of launch decision for launch environments, felt that the concerns expressed by Thiokol and their own SRB engineers were overly conservative based upon prior success. It didn’t help that they didn’t have a specific temperature launch constraint and had launched the shuttle at temperatures below the qualification range of the SRB. Basically, the management talked themselves into a low risk assessment when the criteria was ambiguous and actual data indicated that the field joint was in an out-of-design condition of partial blow-by and o-ring erosion.
This is correct and the actual numbers are 0.988 reliability for solids and 0.975 for liquids on a loss of mission basis in the period 1980-1999.
Those don’t sound very different but looked at as probability of failure, liquids failed at about twice the rate of solids. I’d still take liquids and a robust escape system for manned spaceflight, though.
Actually, three of the four N1 failures (out of four launches) are said to have been in the liquid propulsion system (customarily you count the tanks and feedlines and related stuff in with the engine). But I wasn’t counting data from foreign countries. They do things differently. Also, the stats I quoted in the post above are for large orbital space launches. They don’t count tactical missiles of which there must be tens of thousands if you go down to the small ones. Again these are not really in the same data set as US orbital vehicles.
Solids have seen some significant accidents during ground processing and even during manufacturing because the propellant is right there in the motor instead of being fueled at the launch site like liquids are. Solid propellant is sensitive to shock, heat, etc. Sometimes it has to be machined or sawed and you can imagine the precautions that have to be taken. Anyway, a couple of SRM ground accidents that come to mind are a Titan segment that got dropped by a crane and went off, and a tactical missile back in the 80s that got set off by static electricity. Both incidents resulted in fatalities and there are numerous others.
