After learning how the solid fuel is protecting the casing of the SRB from the heat of the combustion in this answer I have to ask this question.

As we know, it's exactly this segmentation of the boosters that led to the Challenger Disaster

Sure, since the boosters were built in Utah transportation of the almost 600 metric ton boosters was an issue. But were those the only concerns (because I think there would have been other solutions to this with lower risk) or were there manufacturability issues or the like as well that led to the fatefull decision to use segmented SRB with field joints?

Sure, the Question for "failure modes of segmented SRB" did not garner many answers but "risk of leaks" seems quite a big issue to address.

  • 4
    $\begingroup$ The decision to use segmented boosters wasn't fateful, the SRBs would have had a fine record if NASA and contractors had fixed a known issue that had been raised as a risk by using a better material in the o-rings. $\endgroup$
    – GdD
    Oct 28, 2022 at 13:12
  • 4
    $\begingroup$ @GdB there were many "if" in this desaster. If NASA and Tyacol Management had listened to engineers and not put pressure on them instead; if they had decided on a geometry of the field joints that was self tightening; if they had deciced to fix the problem when it was discovered... but at to root are thos field joints due to the segmented boosters imo. $\endgroup$
    – TrySCE2AUX
    Oct 28, 2022 at 13:18
  • 7
    $\begingroup$ You are making the assumption that a single segment would have been safer, I am not convinced you are right on that, or whether it was even possible to do and make it reusable. Hopefully the answers will address that. $\endgroup$
    – GdD
    Oct 28, 2022 at 13:24

3 Answers 3


Four companies submitted bids to build the solid rocket boosters for the Space Transportation System based on the requirements provided by the Space Shuttle Program.

Three of the designs were segmented cases to expedite transport of the boosters from the factory. However the fourth bidder Aerojet submitted a proposal for an unsegmented case.

The Aerojet proposal was rejected specifically because of its case design.

The strength of the case was found inadequate for the prelaunch bending moment loads and was not designed with an adequate safety factor for water impact loads

Of the four proposals, the ones selected for final consideration were those of Thiokol and Lockheed (both segmented case designs). Lockheed was preferred on technical matters, but Thiokol was preferred on management and cost.

There was controversy over the selection of Thiokol over Lockheed, especially because the NASA administrator was from Utah (home of Thiokol) as was a powerful senator who chaired the Senate space committee. Lockheed unsuccessfully protested the award.

However, note that this controversy was between two segmented designs.

Bottom line, segmented cases were used in the design because most of the solid rocket motor manufacturers of the time felt that such a design would meet the requirements best. The contract managers at NASA agreed and selected one of the segmented designs.

Reference: Development of the Space Shuttle, Heppenheimer, pp 71-78

  • 2
    $\begingroup$ is there a good description with images of the Lockheed segment joint design? I'd really like to understand how the Thiokol/Lockheed design approaches differed. $\endgroup$
    – BradV
    Oct 30, 2022 at 14:38
  • 1
    $\begingroup$ @BradV sounds like a great question to ask on the site. I'm not in a position to research it now, but IIRC Lockheed had a superior joint design. A lot of the proposal documents are on ntrs. $\endgroup$ Oct 30, 2022 at 15:09
  • 1
    $\begingroup$ Didn't the segmented boosters fail water impact anyways? $\endgroup$ Oct 31, 2022 at 19:11

As alluded to by Organic Marble, transport from factory to assembly is/was a major issue. The SRB segments are transported mainly by rail flatcar and so have to fit within the standard dimensional envelope of the US freight railroads to insure they will fit through any tunnels, bridges and or multiple parallel track areas between the shipping and receiving points.


Railroad historian here. The apocryphal Roman chariot story actually has a lot of truth to it. To summarize: Roman chariots had a particular, standardized wheel track. This was designed on the basis of the width of the rear of the two horses pulling the chariot. Since these would tend to rut roads (sometimes accounted for in road design in what is arguably the first "rail" roads).... other wagons found they needed to near-match, or the side forces of the ruts would tend to break wheels or axles. So this became a quasi-standard, in turn defining wagon widths.

Which then flowed into wagon design on early tramways, which informed early Railways particularly George Stephenson's successful efforts. There had to be a "standard" track gage, as per frickin' usual, the "popular" beat the "technically superior" (6' was a serious contender and even 7'/2.1m was tried). But here we are at 56.5" or 1435mm.

Loading gage (width/height) was the issue

Of course, track gage is barely the point; it's Loading Gauge that matters, and one does not dictate the other; witness the shockingly small London Tube loading gauges compared to the American West "Plate H Plus" sizes, that handled double-stack containers with very little tunnel modification. But still, track gauge defines a practical limit to Loading Gauge. For any speed, tracks must be superelevated (banked) in curves, and a train must also be able to stop and then handle starting drawbar force in those curves without falling inward, which does in fact happen when an engineer is not careful.

The selection of Thiokol in Utah was another case of popularity beating common sense. Anyone along a coast or navigable river could have shipped a complete 590-ton booster by barge, which would mean field assembly would not be a design requirement. Sites along inland flatlands could have at least sidestepped the diameter issue, by shipping via tunnel-free rail either to a marine terminal or simply rail following along the Missouri, Mississippi or Gulf Coast without a tunnel. Utah, however, is a mountainous bastion. So segmentation and rail tunnels were in play.

But loading gage controlled length due to curves

The segments are quite short for rail cars, which leads into a characteristic of curved tunnels. Curved tunnels must be wider to accommodate both inward swing of long-between-bogeys cars, and outward swing of the tails of long-overhang cars. The largest car generally in circulation is the 89' auto racks or hi-cube boxcars. Since shorter cars have less chord, this allows extra diameter at the cost of less length per segment, which means more segment joints. SRB segments are - well, I couldn't find a positive reference, but a bit shorter than an old 40-foot boxcar (those may be 50's).

Further, in the mechanical age, UPRR was overhauled with generous tunnel clearances, as were most in the west. They are larger than tunnels in the east. And the UPRR routing from Ogden to Nebraska along I-84 and I-80 has relatively few tunnels, and none needed beyond that point.

But yes, it's fair to say the Thiokol booster design was controlled by loading gage (particularly due to curves necessitating short loads) which necessitated the number of segment joints... and that was a bit influenced by - back to the start - a couple of horse's asses.

Namely the guy who picked 56-1/2" rail gauge over wider options, and the guy who really wanted the boosters made in Utah :)


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