SUMMARY OF CASING JOINT DESIGN

¶Issue

¶Why did the aft field joint between the steel containers that hold the Solid Rocket Motor propellant fail to contain the burning gases of the propellant during lift-off and flight operations?

¶John Wark, "Bitter Freeze is Expected to Clobber State Tuesday," The Orlando Sentinel,

¶Jan. 26, 1986,p. B-3.

¶Findings

The design of the field joint was unsatisfactory and could not reliably contain the burning propellant gases under the range of operating conditions to be expected during the lift-off and flight phases.
The O-ring materials and putty used in the design of the joint were unsatisfactory as used on the Shuttle, particularly during the winter months. Furthermore, neither NASA nor its contractor,

¶Morton Thiokol, can adequately control the quality or consistency of these kinds of materials, which are made from recipes known only by the manufacturer and which can be changed without certi- ' fication and approval.

¶Recommendations

NASA should write and issue a new and more accurate performance specification which would cover the full range of thermal and structural requirements for the Solid Rocket Motors, with an adequate factor of safety for unusually low temperatures.
The Committee concurs with the Rogers Commission Report

¶Recommendations on new joint design, but believes it is more a p propriate to be more explicit in identifying the weaknesses in the joint design that need correction.

The field joints of the Solid Rocket Motors should be redesigned to account for the following features while providing a significant factor of safety:

a. Movement in the joint; b. Proper spacing between tang and clevis; c. Seals made to withstand high and low temperatures under all dynamic thermal and structural loadings; d. Adequate sealing without the use of putty; e. Protection against insulation debonding and propellant cracking.

¶Discussion

¶This section is a summary of Section VII, Casing Joint Design.

¶For details and substantiation of the statements made in this summary, refer to Section VII.

¶The evidence, consisting of recovered pieces of the right Solid

¶Rocket Motor casings, photographs of smoke and flame emanating from the right Solid Rocket Motor and telemetry data transmitted from STS 51-L back to Mission Control at the Johnson Space

¶Center verify the failure of the aft field joint of the motor.

¶As mentioned earlier, NASA's performance specifications did not anticipate operations at temperatures below 31 degrees, a temperature that might occur in Florida during the winter months. The design of the joint was unsatisfactory to provide for the low temperatures or water in the joints that existed on January 28. While it was based on an existing similar rocket casing joint design that had been successful, the design was changed to accommodate the manufacturing requirements of the larger sized shuttle rocket motors. There were even some features of the revised design that indicated the changes were an improvement. It was easier to assemble in the field and it had a second O-ring. The designers thought if the first O-ring failed, the second would surely hold the propellant gases.

¶The casing joints, as described in the Introduction, have to withstand various structural loads, which change dramatically as the shuttle is assembled, through launch operations, separation of the Solid Rocket Booster and retrieval from the ocean. The joint is dynamic; the components move under these loads. The loads carried by the aft field joint are different from those carried by other joints. The design, based on these loads and 24 successful missions, appeared satisfactory.

¶One of the loads, however, that of the propellant gas pressure, was not adequately accommodated. The zinc chromate putty, intended to protect the O-rings from this high temperature and relatively high pressure gas, frequently failed and permitted the gas to erode the primary O-rings.

¶Instead of redesigning the joint, NASA and Thiokol persisted in trying to fix the problem by changing leak-test pressures, changing the size of the O-rings, and trying to control proper spacing between the tang and clevis where the O-rings were located.

¶Complicating this problem, two of the materials used in the joint, the putty and the fluorocarbon elastomer O-rings, were not suited to the task of containing the propellant gas under the full span of Shuttle operating conditions. The behavior of the fluorocarbon elastomer O-rings was something of a mystery to NASA and its contractor. The material was "proprietary," meaning that the con- stituents used were known only to the manufacturer. Fluorocar- bons are expensive, so fillers are frequently added to reduce the cost of the material. These materials behave unlike most other materials. The particular material used in the manufacture of the shuttle O-rings was the wrong material to use at low temperatures. Nitrile or silicon based materials would have demonstrated better performance characteristics.

¶It became necessary to find a new putty when the original supplier, Fuller O'Brien, stopped making it because it contained asbestos. The characteristics of the new putty changed substantially in response to the quantity of water in the air and it was difficult to apply in both the dry climate of Utah and the dampness of Florida. Its performance in use was highly unpredictable. Again, NASA and its contractor tried to make up for the unsatisfactory material by storing it under refrigeration prior to application in Florida.

¶After ignition of the solid propellant in the SRM, It was learned that the O-ring could be seated by the motor's gas pressure yet still suffer erosion as the hot gases came in contact with it. As mentioned, O-ring erosion was noted after various flights and tests. Also seen was damage given the name "blow-by", a condition where erosion was not necessarily present but where there was evidence that the propellant gas had bypassed the primary O-ring. But rather than identify this condition as a joint that didn't seal, that is, a joint that had already failed, NASA elected to regard a certain degree of erosion or blow-by as "acceptable." To make matters worse, confidence was mistakenly obtained from a mathemati33NASAa primary concern was having a very durable material with excellent high temperature performance characteristics.

¶cal model which suggested that if the erosion did not exceed a specific depth, the O-ring would still seal that joint. In cases where the erosion did exceed the maximum predicted by the model, NASA expanded its experience base to cover this increased damage.

¶As the joint seals continued to exhibit erosion or blow-by or both, more research illustrated the importance of maintaining proper gap spacing between the tang part of the joint and the O-ring face of the inner clevis leg. Too little space, and the O-rings would not seal. Too much space, and again the seals would fail. Since the joint opens, or "rotates," when the Solid Rocket Motor is ignited, maintaining proper spacing was difficult if not impossible. The maintenance of such close tolerances in spacing, on the order of 20 thousandths of an inch, while joining 300,000 lb. segments that have been bent during shipment, was not sufficiently provided for in the design. Months passed until, in 1985, engineers at NASA recognized that the design was unsatisfactory. In fact, NASA had written to several other contractors soliciting help with the joint problems. Unfortunately, in the quest to meet schedule and budget, the warnings of the engineers were not heeded.

¶Based on the above conditions and the evidence, the Committee has endeavored to determine the way in which the joint failed; recognizing that such a determination is difficult, if not impossible, to make with 100% certainty.

¶The following is the most probable sequence of the joint failure:

The failure occured in the lower assembly joint near a strut that connects the Solid Rocket Booster to the External Tank.
At that location, the spacing between the two casings was too small to facilitate a tight seal.
Also, at that location, there probably existed a hole through the insulating putty, which would act as a conduit concentrating the hot propellant gas on the primary O-ring.
The freezing temperatures reduced the capability of the 0- rings to seal. Worse, at this particular location, near the connecting strut, the joint was made even colder by the further loss of heat caused by the direct connection to the liquid hydrogen fuel, at 423 degrees below zero, in the external tank.
When the Solid Rocket Motors were ignited, the pressure from the motor changed the spacing between the casings. Among other effects, this can prevent the secondary O-ring from sealing.
Seven inches of rain fell while the shuttle was being prepared for launch. Water very likely penetrated the joints and froze. Ice in the joints could have dislodged the secondary O-ring even if the change in spacing, coupled with a cold and stiff O-ring, did not.
Smoke at ignition occurred at a location near the connecting strut to the external tank. At that location, the primary O-ring was either unseated or eroded and the secondary O-ring was unseated.
The primary O-ring was sealed at other locations around the motor casings.
The breach in the primary O-ring clogged with burned char and aluminum oxide from the -propellant- in less than 3 seconds, causing the smoke to stop.
At 37 seconds,⁴⁵⁴⁵ Ibid,. Volume 11, p. K-23. seconds and 58 seconds into the flight, the Space Shuttle encountered heavy turbulence, which forced the steering controls to cycle through changes more severe than previous flights.

After throttling back to 65% power as planned, at 57 seconds, power was increased to 104%.
The combined effect of the turbulence and the increase in power caused the material which clogged the joint to break free, reopening the joint.
A flame from the right Solid Rocket Motor was seen at the location near the connecting strut.
This flame burned through the external tank and caused the destruction of the shuttle.

¶Since the technical faults in the joint design must be corrected if safe shuttle flight is to resume, this subject has been discussed in more detail in Section VII.

TESTING A N D CERTIFICATION

¶Discussion

¶In developing the Solid Rocket Motor, Thiokol concentrated most of their efforts and concerns on the proper design and performance of the propellant. There is no question that this is where the emphasis on safety and performance is required. The propellant is a high performance material, dangerous to manufacture and handle and which must be prepared to the highest quality standards. Consequently, testing and certification of the propellants, as well as its performance, was carefully controlled. This does not mean that the design of the casings was ignored. Considerable attention was paid to the design of the casings because they were larger than seen on any previous Solid Rocket Motor, because this Solid Rocket Motor would be used on a manned flight system, and because these particular motors would be brought back, refurbished and reused. Given this background, the testing of the joint was included in static firing tests. While there were no special tests conducted to confirm and certify the joint as a separate item, analysis was performed to assure that the joint was adequate. Later, during the operation of the Solid Rocket Motor, it was discovered that the performance of the joint was unsatisfactory.

MANUFACTURING

¶The Solid Rocket Motor is 126 feet long and 12 feet in diameter. The propellant weighs 1.9 million pounds and the average thrust is 2.3 million pounds. Fifty of these motors have been produced. The segmented Solid Rocket Motor case is roll formed from D6AC steel. The case is weld-free and consists of eleven segments. The propellant is made in batches a t 135 degrees F and it takes 40 to 43 of these batches to load one casting segment. One segment includes two steel cases which are joined in the factory. The content and quality of the materials usea to make the propellant is inspected prior to mixing. The motor is designed for a short burn time (122 seconds) and therefore has a high mass flow which requires a large burning surface. In manufacture, either new steel casings or previously used casings are employed. The first step is to apply the rubber insulation liner around the inside of the casings. The insulation is removed from a roll and spread around the inside of the casings with special tooling. After application it is cured in place in an autoclave. After the casings have been insulated, they are placed in a casting pit. The propellant is then poured into the casings under vacuum. The propellant is then cured and the casings are removed from the pit. There is no indication that there were any manufacturing defects that contributed to the loss of the Challenger.