28
$\begingroup$

All that I know about the Challenger tragedy was that some o-rings failed in the reusable solid rocket boosters (SRBs) due to the low temperatures that day. I'm curious if such an accident would have happened if the SRBs didn't have to be reusable.

For example, maybe whatever coupling that used the o-rings could have been solid or sealed in another way if the SRBs weren't intended to be retrieved and reused. Or, maybe some other specific implementation detail of the SRBs or the SRBs inherently weren't ideal but the goal of re-usability trumped that.

$\endgroup$
  • $\begingroup$ AFAIK the SRBs had to be manufactured in segments, thus the O-rings between each piece. $\endgroup$ – Jerard Puckett Jul 19 '16 at 21:38
  • 2
    $\begingroup$ A more certain avenue to a safe flight would have been found in either: 1) a more flexible schedule (the teacher would have been broadcasting to empty class rooms if they delayed it). Or 2) NASA staff not pressuring the engineers that made them, for an 'OK to fly' approval. - Still, '20/20 hindsight' & all that.. $\endgroup$ – Andrew Thompson Jul 19 '16 at 22:09
  • 3
    $\begingroup$ The implementation detail of the SRB that as inherently a bad idea was building it out of state (ie not in Florida), and thus requiring them to glue it together in Florida (ie weak spots). $\endgroup$ – Aron Jul 20 '16 at 8:32
  • $\begingroup$ check out the book 'no downlink' $\endgroup$ – Craig Constantine Jul 20 '16 at 16:47
  • $\begingroup$ @JerardPuckett: They could still have been manufactured in segments and then welded together before filling them with propellant. $\endgroup$ – Sean Jun 10 '18 at 18:06
61
$\begingroup$

The O-rings and the temperature were only the last in the long chain of blunders, and that had little to nothing in common with reusability.

The construction of the SRBs wouldn't be much different with no reusability in mind. Indeed, SRBs of very similar design are to be used in SLS, and they are not intended to be reusable.

About the most important factor responsible for the disaster was misunderstanding of how the O-rings were meant to function.

The segments of SRBs were a consequence of the technological process; the boosters were manufacturer that way, regardless of the reusability needs. The segments would be connected with seals consisting of the O-rings, and ablative putty. The putty would act as a sealant keeping the pressure in, keeping the flames from escaping, and providing thermal isolation to the O-rings from the flames. The O-rings were to keep the putty from getting squeezed out by the extreme pressures of combustion. They were never meant to come into contact with flames and keep them from escaping, to any degree.

The putty appeared to be not efficient enough - all of it would burn away during the SRB burn. Then - the O-rings - not meant to withstand any flames - would still withstand them to such a degree, that less than half of them would burn through. And that fact - misunderstanding the fact that the putty burning through entirely was already a critical failure, and treating the degree of ablation of the O-rings as indicative of performance of the seal - allowed continued use of the faulty design.

The O-rings, not ever being meant to operate as flame blocker, not tested or prepared for such a role, would operate within specs for their designated task, and entirely out of specs for the task they were not meant to. It was pure dumb luck they would perform as flame blocker at all, and with changing weather conditions of course their behavior as flame blocker changed - nobody ever engineered them not to!

In fact - reusability - and in particular, recovery of the boosters - was a factor that could have prevented the disaster: the damage to the O-rings might pass unnoticed otherwise. Thanks to the boosters being recoverable, the damage was noticed, analyzed, and had the right conclusion been drawn, it would lead to redesign and fixing the error, e.g. replacing the putty with one that ablates slower.

Unfortunately - despite that - the problem was ignored, the damage accepted as "within norm".

So, in the end, the conclusion is pretty much opposite to the one you implied: Reusability is more likely to prevent disasters, than promote them; it allows to notice and analyze any problems that occur and would be missed in non-recoverable rockets until changing external conditions lead to a disaster. It just requires a thorough insight into the intended workings of the rocket, to be able to draw the correct conclusions from the observations. Without reusability we may lack the observations to draw conclusions from.

$\endgroup$
  • 7
    $\begingroup$ @tau: Reusability is the secondary driving force. The primary force is testing. These things should have been detected and mitigated in the ground tests phase, even before the first launch. Still, not everything can be tested on the ground, not all conditions can be simulated, and not all problems are repeatable enough for tests to reveal them. $\endgroup$ – SF. Jul 20 '16 at 0:56
  • 19
    $\begingroup$ @RussellBorogove: That was again, one of the last nails to the coffin. The problem was known for a long time, possibly years. Basing the launch decision upon weather here is akin to having an airplane with its wings glued to the hull with chewing gum and deciding whether to cancel the flight today because the sunlight may soften the chewing gum. And after the disaster analyzing why the chewing gum failed, instead of asking why the wings weren't structurally welded to the plane frame. $\endgroup$ – SF. Jul 20 '16 at 6:54
  • 6
    $\begingroup$ You really have to understand that, more then any thing else, the people in charge wanted to launch. They would continuously ignore the recommendations of their engineering staff in order to keep a schedule. The "O-Ring problem" was known for years and was deemed not important enough to pay for a reworking. "Meh the last time it didn't blow up so it should be good this time, why do I need to shell out millions for a washer just in case. Launch! Launch! Launch!" This failure was a direct result of a systemic failure at NASA (and others). The cold weather played it's part, but it was a tiny bit $\endgroup$ – coteyr Jul 20 '16 at 13:05
  • 11
    $\begingroup$ @coteyr: Yes, the original sin was the organization culture of NASA. That was what allowed this misinterpretation to persist - there were people, who realized this is a critical failure, but they just couldn't "get through" to the management, which was completely convinced 50% of O-rings remained meant 2x safety factor (while any O-ring damage meant in fact zero safety factor.) $\endgroup$ – SF. Jul 20 '16 at 13:12
  • 2
    $\begingroup$ @tau To re-use the boosters, they needed to be disassembled, thoroughly cleaned, inspected and otherwise refurbished. SF. is saying that during that process, the O-ring failures were discovered. This happy accident was a golden opportunity to correct a problem that design and testing had missed. A non-reusable assembly would not have given NASA this second chance to discover the error. $\endgroup$ – JS. Jul 21 '16 at 21:16
24
$\begingroup$

Not reusability per se, but design and politics. The competitor to build the SRBs for the Shuttle program was Aerojet, and their design was monolithic. In other words, no segments. Aerojet was actually selected as the winning contractor but the NASA administrator Fletcher overrode this and gave it to Morton Thiokol. Aerojet had already built an awe-inspiring test facility in Florida.

By the way, Fletcher was from Utah.

It will never be known whether Aerojet's booster design would have had fatal flaws as well, but it would not have had the actual flaw that doomed Challenger and the crew.

Sources

Texas Space Grant Consortium (Describes Fletcher's overturn of contract award and subsequent controversy)

philly.com (Describes how the Aerojet design was monolithic and presumably safer)

$\endgroup$
  • 1
    $\begingroup$ In addition to what Organic Marble said about Aerojet politics played another role--a juicy piece of pork like this had to be handed out to as many powerful representatives as possible. Practical considerations like how to transport the boosters were secondary to this. The Morton Thiokol booster had to be in segments because there was no way to transport a whole SRB from their factory to the Cape. $\endgroup$ – Loren Pechtel Jul 20 '16 at 3:44
  • 1
    $\begingroup$ The Aerojet boosters would have been transported by barge to the Cape. $\endgroup$ – Organic Marble Jul 20 '16 at 13:05
  • $\begingroup$ According to this: oldrocketforum.com/printthread.php?t=5656&pp=40 the Aerojet booster showed serious problems during tests, traced back to the more variable climate in Florida, whereas Morton Thiokol in Utah could perform more consistent casting of the propellant. $\endgroup$ – Michael Borgwardt Jul 22 '16 at 15:42
  • 1
    $\begingroup$ @MichaelBorgwardt I believe that was the 260 inch motor which was intended as a 1st stage replacement for the Saturn. The last test of a subscale version had a serious failure. I don't think they got very far on Shuttle motors. But it's been a long time. $\endgroup$ – Organic Marble Jul 22 '16 at 16:05
15
$\begingroup$

Technically, it was the O-rings, but that was not the root cause. The root cause was a poor culture at NASA. The technical solution was easy: Don't launch on that cold of a day. In Florida, even in late January, all you have to do is wait a day or two and temperatures will once again be balmy. Technically savvy people knew well of the O-ring problem in cold weather. They tried to escalate these concerns but were completely rebuffed. NASA had a schedule to meet.

Poor culture at NASA was also responsible for the Apollo 1 fire, the Columbia disaster (two references), Hubble's messed-up optics, and lots of lesser problems.

$\endgroup$
  • 8
    $\begingroup$ Avoiding cold weather was not a complete solution. As SF's answer mentions, the O-rings should not have been directly exposed to combustion at all (and were not, after the joint redesign). $\endgroup$ – Russell Borogove Jul 20 '16 at 5:03
  • 2
    $\begingroup$ @RussellBorogove Nevertheless, the point is that the disaster was not the result of a single problem, but rather a series of problems and failures that each contributed to the final outcome. Engineering projects always have design issues that need to be worked out - not all of them end in disaster, however. Cultural, political, and managerial problems were equally to blame here - the O-ring was certainly the point of failure, but it's crazy to say that the o-ring alone was the cause of the disaster. Lots of other things had to go wrong to get to that point in the first place. $\endgroup$ – J... Jul 20 '16 at 15:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.