O-RING W I L L NOT 1 SEAL

O-RING W I L L NOT 1 SEAL

187

3. However, there could have been one or more than one blow- hole through the zinc chromate putty before ignition. One such blowhole could have been made at the 300 degree location either during the leak check at 200 psi prior to the seating of the primary O-ring or prior to the leak check when the casings were brought together. If so, there was a high probability that the primary 0- ring eroded at this point. The phenomena of blowing holes in the putty had been observed many times at post-flight dismantlement and had dramatically increased when the procedures were changed to increase the test pressure to 200 psi. Additionally, the Randolph putty had been found unsatisfactory on numerous occasions.2

4. Upon ignition of the Solid Rocket Motor, this blowhole would have facilitated and concentrated the hot propellant gas flame on the primary O-ring, and possibly the secondary O-ring as well. Alignment of O-ring erosion with the location of blowholes had been observed on numerous occasions.3

5. Between the time the casings were assembled and the launch, the secondary O-ring was unseated from its previously sealed posi- tion. The fact that it had been sealed has been verified by the pres- sure check made 28 days before when the casings were joined. Either the secondary O-ring was unseated by joint rotation coupled with O-ring stiffness or by the formation of ice in the joint.

¶During the intervening period, as the Shuttle stood on Pad 39B waiting for the launch,⁷⁷ bid. inches of rain had fallen and some could have easily penetrated the joints. The access of rain water into the joints was proved when STS-9 was disassembled and water poured out of the assembly pin holes. In tests conducted after the accident, it was confirmed that the water in the aft field joint would have turned to ice, and that the ice could have dislodged the secondary O-ring, pushing it upstream into a non-sealed position. In this posi- tion, it is doubtful that the secondary O-ring could have sealed at ignition.

6. One of the three Solid Rocket Booster to External Tank aft at- tachment struts is also connected at the 300 degree location, just a few inches below the aft field joint. As the Space Shuttle system stood on the launch platform at Pad B on January 28, the large External Tank was gradually filled with liquid hydrogen and liquid oxygen. Liquid hydrogen, at a temperature of 423 deg F below zero, and liquid oxygen, at a temperature of 297 deg F below zero, caused the tank to contract as it was filled. Since the Solid Rocket Boosters are firmly bolted to the launch platform, a lateral force of approximately 190,000 pounds pulled sideways on the aft attach- ment strut and the Solid Rocket Motor casing, including the joint that failed.4 Refueling of the tank was accomplished early on the morning of January 28.

¶At ignition, the 190,000-pound force was instantly released when the SRB hold-down bolts were blown loose. For the next two and a half seconds the right Solid Rocket Motor field joints experienced a 3 cycle per second vibratory load caused by the sudden release of

¶
• NASA, MSFC Memo, Miller to Horton, April 12, 1984.

¶Thiokol, "Erosion of SRM Preasure Seals," TWR 15160, Chart A-9, August 19, 1985 "Seal damage alwa s has associated putt blowhole."⁴⁴ Rogers Commission Report, Volume 11, p. H-1. NASA, &FC, "51-L Analysis &erview," April 25,1986, p. H-203.

188

¶the lateral force.6 The ignition pressure increased the joint spacing. Also, the flow of motor gases through the blowhole at the 300 degree location could have resulted in damage to the primary 0- ring. The evidence of smoke at the 300 degree location is unlikely without O-ring damage.

7. Smoke at launch, clearly visible in the photographs, stopped at 2.7 seconds when the vibratory load damped out and the joint sealed. The sealing of the breach at the 300 degree location was made possible by blockage from burned material, probably consist- ing of a mixture of insulation and aluminum oxide. Post-accident tests performed by Morton Thiokol proved that aluminum oxide could have successfully plugged the joint at 2.7 seconds. While the smoke at ignition appeared to be intermittent, that appearance was probably a result of air and main engine exhaust currents.

8. At T+37 seconds into the flight, the Shuttle encountered wind gust loads in conjunction with planned maneuvers. Components of these gust and maneuvering loads were transmitted to the Solid Rocket Booster through the External Tank attachment strut. Based on the prescence of smoke at liftoff, these forces were transmitted to a joint already weakened by erosion and heat damage.

9. At 43 seconds into the flight, the main engines throttled back as the Shuttle reached "Max q" (maximum dynamic pressure). Four seconds later, the main engines had throttled up to 104% power and the geometry of the Solid Rocket Motor propellants had increased thrust. At this point, the motor pressure increased to 609 psi.

¶Additional structural loads resulted from turbulence. Flight 51-L experienced the most severe turbulence of any Shuttle flight and, although the loads were within the allowable design limits, those design limits did not consider a joint that had already failed.s It is unknown how much the combined effect of wind gust loads, ma- neuvering loads and an increase in thrust contributed to the acci- dent. But the combined effects of these forces could have dislodged the burned material at the previously breached section of the joint.

10. Shortly after the vehicle was loaded by these turbulent forces, at T+58 seconds, a flame appeared from the same general region where the puffs of smoke had been seen. But, this time the joint was continuously breached by the burning propellant gases. In a little over two seconds, the flame had grown and acted as a blowtorch to burn through the hydrogen tank. The appearance of the flame at this time is also indicative of a damaged primary 0- ring and failure of the secondary O-ring to seal, for reasons ex- plained in the Critical Items List dated December 17, 1982.sa

¶The telemetry data, photographs and cockpit voice recordings support evidence of turbulent conditions and the manner in which the Shuttle failed.

¶6 The joint waa designed to accommodate these loads. 0 NASA, MSFC, "51-L Analysis Overview, STS 51-L-Wind Shears, April 25, 1986, p. H-597. 0' NASA, "SRB Critical Items List," December 17,1982, page A-6A, sheet 1.

189

¶Time ( m i n x ) Crew position Crew wmment

¶T+19 .......................................................................... PLT................................... Lwks like we've got a lotta wind here today. T+20 ............................................ .. CDR .................................. Yeah.

. CDR .................................. It's a little hard to see out my window here. . PLT................................... There's ten thousand feet and Mach point five. ..................................... (Garble)

¶T +35 .......................................................................... CDR .....................

..................................... [High thrust vector control (steering) activity noted. This was caused by upper atmoshpere wind gusts and planned maneuvers.]

¶T+41 ......................

¶.................................................... CDR ................. OK, we're throttling down.

¶T+58 ........................ ......................................... [Vehicle loaded by dynamic pressures.] T+58 ........................

where the puffs of smoke had been Seen earlier.]

¶T+ 59 .......................................................................... CDR .................................. Roger. T+60 .......................................................................... PLT T+60 ................................ .................................... [Right SRM internal pressure began to diverge from that of left SRM] T+ 61 .............................................................................................................. [Well-defined plume was deflected indicating the

plume had burned through the liquid hydro- gen tank structure.]

¶T+62 .......................................................................... PLT................................... Thirty-five thousand going through one point fiw T+64.7 ........................................................................................................... [ti T+66.8 ..................................................... . Reading four eighty six on mine. 66.800 ........................................................ . [Leak confirmed when hydrogem tank leak

pressurization system was unable to maintain normal pressurization rate.]

¶T+67 .......................................................................... PLT ............ Yep, that's what I've got, tw. T+ 70 .......................................................................... CDR ........... Roger, go at throttle up. T+72.2 .................................................................................... [Right Solid Racket Booster motion differed

from Orbiter and left Solid Rocket Booster, indicating failure of lower attachment struc- ture.]

¶T+72.6 ......................................................... [Liquid hydrogen tank pressure fell. Leak was growing rapidly.] T+73 .............................................................................................................. [Liquid hydrogen and liquid oxygen pressure to main engines showed significant drop.] T+73 .......................................................................... PLT................................... Uh Oh T+73.1 ........................................................................................................... [Circumferential white pattern around the Exter-

nal Tank aft bulkhead suggested liquid hydro- gen tank structure failure.]

¶1+73.1 ........................................................................................................... [Vapor observed at inter-tank which was indica-

tive of the liquid oxygen tank failing. Liquid oxygen then 0bserved.1~

¶1 (CDR) Commander W, (PLT) Pilot Smith, (MS 1) Mission Specialist Onizuka. (MS 2) Mission S ialisl Resnik.⁸⁸ Ibid. Ibid. ' 0 bid. I 1bid. Morton Thiokol, "Program Plan, Protection of Space Shuttle SRM Primary Seals," TWR- 14359,May 4, 1984. s Ibid. Discussions with Allan McDonald and Carver Kennedy, Thiokol (Wasatch Operations), Brigham City, Utah,, September 4, 1986. NASA, D.M. Germany, STS 511 Incident Investigation. Integrated Events Time line, Johnson Space c h t Center, June 4, 1986, as modified.

¶(c) Problems Discovered

¶The design of the joint was based on the successful design of the joints used on the Titan I11 booster r ~ c k e t . ~

¶That design was simi- lar except that the tang pointed upward, instead of down, and the clevis pointed downward, instead of up, as in the case of the Shut- tle booster. Another difference was that the design of the Shuttle joint included two O-rings instead of one as provided for in the

¶NASA, "SRB Critical Items List," December 17,1982, p. A-6A, Sheet 1.

190

¶Titan design. But, the most important difference was the use of putty in the Shuttle design. While the Titan employed the NBR in- sulation to close the gap between segments, the Shuttle design called for filling a gap between insulation with putty.

¶The Shuttle design was changed to accommodate manufacturing constraints.'O The Shuttle booster is larger,¹⁴⁶¹⁴⁶ bid., p. 73. inches in diameter as compared to 120 inches for the Titan. As a result of its larger size, the Shuttle booster uses more steel. While this requirement for more steel had no impact on other booster components, it did have an impact on the joint design. The maximum billet size (a piece of metal made from an ingot) commercially available to man- ufacture the large, one-piece, weld free forward dome with an inte- gral forward skirt tang was less than that needed for the Shuttle Solid Rocket Boosters. However, it was found that by turning the casings upside down, there would be just enough metal to manufac- ture a forward dome because that component would then only have to incorporate the single joint element, the tang, instead of the double joint element, the clevis.

¶It is good engineering practice to design products to accommo- date manufacturing tooling capabilities and methods. Furthermore, with the clevis facing up and the tang down, field assembly at the Kennedy Space Center was simplified. Combined with the extra 0- ring, the design change appeared reasonable. But it is also good en- gineering practice to accommodate all the forces and conditions that the product must perform under during its useful life. The design of the Shuttle Solid Rocket Motor, as opposed to the Titan, had to provide for reuse of the propellant casings, including the wearing of joint surfaces and distortion of the case in handling and shipment. It had to accommodate heavier propellant loads. The design was more susceptible to water entry during storms. And, most significant, the design had to accommodate a combination of dynamic structural loads significantly different than those encoun- tered by the Titan.

¶It is always a simple task to find fault with someone else's work; especially after an accident occurs. It is quite another matter to originate the work and produce a useful product.

¶The joint design provided a direct path between the combustion chamber, consisting of an annulus with propellant surrounding it, and the outside of the steel motor casings. That path was sealed with putty and two circular fluorocarbon elastomer (rubber-like) bands called O-rings. While O-rings are frequently used to retain pressures much higher than those present in the Shuttle Solid Rocket Motor, thermal and structural forces acting on the Shuttle joints are formidable. These joints must carry and transfer these loads between the casings.

¶Another essential ingredient of good engineering practice is to use material suited to the function. Some O-rings can withstand high temperatures. But "all . . . elastomers become brittle at low [temperatures]. . . . Elastomers, like natural rubber, nitrile rubber, and Viton A . . . that become brittle at low [temperature]

¶' 0 Staff discussion with E.G. Dorsey, Thiokol Waeatch Operations, Brigham City, Utah, Sep tember 4, 1986.

191

¶can be used for static seal gaskets when highly compressed at room temperature prior to cooling."l1

¶But the Shuttle's O-rings were not used in a "static" system as evidenced by the variations in gap spacing between the tang and clevis. Nor would they always be highly compressed at room tem- perature prior to the cooling. Furthermore, the O-rings could not withstand burning propellant temperatures in the range of 5800" F. The design of the joint therefore provided for putty to insulate the O-rings from the burning gases.

¶This putty did not always perform as had been expected, and evi- dence of hot gas passing the putty and getting to the first, or pri- mary, O-ring along the path to the outside of the rocket chamber was discovered. Once the putty was breached, the joint was not working as it had been designed. This failure, although recognized by NASA and its contractor Morton Thiokol, was neglected on March 8, 1984 when they chose to accept an "allowable degree of erosion,'' which meant there was an allowable percentage of fail- ure.

¶O-rings become effective (are seated) when pressure is applied to them as they sit in a groove provided to house them.

¶One question that the design was intended to answer was wheth- er or not the O-ring was seated properly in its groove. An opening, with a fitting much like a valve stem on a tire, was provided to allow pressure testing between two such O-rings, the primary and the secondary.

¶But this design did not always answer the question: was the pri- mary O-ring seated? Did it seal or not? Notice how the primary 0- ring in Figure VII-1 (p. 176) is forced upward (shown by the single arrow). That is opposite to the normal direction that the propellant pressure acts (notice the double arrows). Even with an acceptable pressure check result, the primary O-ring would still be unseated for a fraction of a second when the motor pressure pushed the 0- ring in the opposite direction from that which took place during the leak check.

¶The second reason the assumption concerning the leak check as "proof of sealing'' could be erroneous was that the primary O-ring did not really have to seat at all if the putty behind it (toward the inside of the case) held the pressure during the leak check. So ear- lier in the program there really was no way to know whether the primary O-ring seated or not.

¶What appeared to be a rather straightforward joint was far from simple. If the primary O-ring did not seat during the leak check, and the pressure test succeeded, then the putty was doing the work of sealing. But it still was not possible to determine from outside the casings whether the putty or the O-ring was holding the pres- sure. But, if the leak check failed, then the O-ring was not seated and there was a blow-hole through the putty.

¶To resolve this concern, NASA and its contractor, Morton Thio- kol, changed the leak-check procedure by increasing the pressure until a pressure of 200 pounds per square inch (psi) was accepted as

¶Theodore Baumeister, Editor, Standard Handbook for Mechanical Engineers, 7th Ed., (New York: McGraw-Hill, 1967). pp. 18-35.¹¹ For the purpose of this report, a procedure is a formal set of instructions designed to guide and assist in the performance of a technical or management function. g8 mid., July 24, 1986, p. 11. 2 Rogers Commission Report, Volume 11, p. H-1.

192

¶the standard. They had ascertained that this was sufficient pres- sure to blow a hole through the putty.13 Then, if the O-ring failed to seat, the pressure would blow a hole through the putty and the test would disclose an unseated O-ring (a failed seal). But if the 0- ring held the higher pressure, the O-ring would still have been seated in the upper position instead of the downward position. That would be contrary to the way the O-ring would have to be seated to contain the propellant pressure during launch of the Shuttle.

¶In summary, there was still no way to verify whether the pri- mary O-ring was seated properly, meaning in the downstream posi- tion after the cases were joined together in the field. In the begin- ning of the development program the concept was that the putty would act somewhat like a "piston in a cylinder" when the propel- lant was ignited. As the chamber pressure built up, the putty was to move downstream and compress the air in the path between it and the primary O-ring. The compressed gas was to seat the O-ring and thereby seal the joint. Besides, even if the primary O-ring didn't seal, surely the secondary O-ring would, since it had already been pressure checked, which verified it was seated in the down- stream position.

¶There was no direct evidence that the primary O-ring was not holding the pressure off the secondary ring until Flight 51-B. That was the first flight when erosion of the secondary O-ring had been observed, even though erosion of the primary O-ring had occurred before.

¶Thiokol had considered the joint design to be Criticality lR,15 meaning that there was redundancy. While the second O-ring was redundant by design, the joint as a whole was still Criticality 1, since if it failed, it would mean the loss of the Shuttle and crew. In other words, there was no backup for the joint.

¶The joint was designed to mate two rocket motor segment cases, one to the other, where the lower edge of the upper case consisted of a tang and the upper edge of the lower case consisted of a clevis. After the tang was inserted into the clevis (which housed the two O-rings), 177 steel pins, each approximately 1 inch in diameter, were inserted from the outside through aligned holes which went through the outer leg of the clevis, the tang and partly into the inner leg of the clevis. The spacing between the inner face of the tang and the mating face of the inner leg of the clevis where the 0- rings were housed was critical to the integrity of the joint because that spacing, in part, determined whether the O-rings could func- tion properly to seal against the propellant gas pressures. Not only was the initial static spacing critical, but maintaining the proper spacing during launch and flight under dynamic structural load- ings was necessary for an effective seal.

¶Upon ignition of the Solid Rocket Motor fuel the opcrating pres- sure increases to 922 psi at 40 degrees F within a little over one- half second (0.648 sec).16 The effect of this pressure increase is to

¶NASA, MSFC, Problem Assessment S stem Record No. A07934, January 23, 1986, p. 6.¹⁵¹⁵ Rogers Commission Report, Volume 11, p. K-27. 1'Rogers Commission Report,Volume 4 p. 1510. 'SCmte H Transcri t, June 18, 1986, 51. 1eMortOn%iokol, &-10212 (CD), Tab, 4-9, npical Propellant Design Data.

193

¶cause the casings to bulge out around their midsections while being constrained by the thicker steel sections at the ends, much like a can of soda after freezing. The casings change shape during the buildup of motor pressure. This bulging has an effect on the joint. As in the case of the frozen soda can, the wall of the casing near the joint is no longer vertical, or perpendicular to the bottom, but angles out to meet the larger diameter in the center of the casing. NASA calls this change in angle at the joint "joint rotation."

¶This joint rotation is a component of an overall spacing problem that includes: changes caused by casing wear and tear experienced during refurbishment; case growth (swelling) from pressurizing the casings; distortion that occurs during shipment of the loaded cas- ings; and the physical handling of the casings during stacking oper- ations.

¶The joint rotation problem was aggravated when the steel cas- ings were made thinner to achieve a reduction in weight and thus an increase in payload. The rotation problem was further aggravat- ed by changing the design of the propellant geometry to achieve greater thrust. This increased the pressure within the casings and thereby increased the "gap opening."l These changes compro- mised the integrity of the joint seals because joint rotation in- creases the spacing (gap) between the tang and the O-ring grooves in the clevis.

¶When the increase in the gap occurs, it can open the O-ring seal, leaving the path from the propellant combustion chamber open to the outside of the casing, except for any blockage by the putty. But, as noted above, the putty frequently has holes blown through it. If there were blowholes in the put,ty, and the original spacing be- tween the metal parts of the joint was such that the joint rotation left open spaces between the O-rings and the tang, then the joint would fail and burning gases would escape to the outside.

¶(d) Joint Behavior

¶In a memo from John Miller to Mr. Eudy of NASA on June 16, 1980, the following statement was made:

STA-1 test data shows that the secondary O-ring can become unseated from the tang due to joint rotation at ap- proximately 40 percent of MEOP [Mean Effective Operat- ing Pressure], and therefore, is not likely to assume a seal- ing position should the above primary seal failure occur. The SRM has never been tested to evaluate the above fail- ure condition, nor has credibility of such a failure been of- ficially declared.

¶In March of 1984 Thiokol had completed its SRM O-ring assem- bly test plan, which was to confirm the O-ring erosion scenario, provide data for heat transfer predictions and establish the effec17 The Light Weight Casings, first used on STS-6, had thinner casing walls than the standard steel casings. Llght we' ht casings permitted fli ht "th heavier pa loads. On STS-8, NASA began using the High Verformance Motor (HP& whlch developed iigher internal pressures while using the light weight casings. The purpose of the HPM was to further increase payload capacit

¶Is N h , "Evaluation of TWR-12690 CD, Test Plan for Space Shuttle SRM Li htweight Cyl- inder Segment Joint Verification, dated June 10,1980",EP 25 (80-701, June 1 6 ,¹¹ For the purpose of this report, a procedure is a formal set of instructions designed to guide and assist in the performance of a technical or management function. g8 mid., July 24, 1986, p. 11. 9 8 , p. 2.

194

¶tiveness of the vacuum putty. The introduction to that plan includ- ed the statement:

¶O-ring seals in rocket motors in general and the Space

¶Shuttle SRMs in particular, can suffer thermal degradation because of exposure to the high temperature motor chamber gases. Although none of the SRM primary 0rings to date have failed to perform their design function, there is some concern because of isolated events which show localized erosion as high as 0.053 inches. The postulated scenario for this thermal degradation effect is a short-time duration impingement of a high energy jet which is induced during ignition pressurization by a combination of voids in the protective vacuum putty and the filling of available free volumes created by the tolerances of mating parts and the O-ring slots. Unfortunately, the overall assembly and the vacuum putty layup does not lend itself to a well-defined geometry for predicting the hot gas flow and associated heat transfer to the O-rings.

¶A subsequent report, dated May 7, 1984, contained a statement:

¶Symptom of failure: a vaccum putty exhibited gas paths located at 319 deg., 338 deg., and 347 deg. Erosion of the primary O-ring occurred at 319 deg. only. The damaged region was approximately 5.6 inches long with a .034 inch maximum depth and involved 136 deg. of the O-ring cross section diameter.20

¶In a memo from Larry Mulloy to Bob Lindstrom, Director, MSFC Shuttle Projects Office, in November of 1984 it was noted:

¶. . . it was determined that shims could be used to make the case joint sufficiently concentric to consistantly achieve a 7.54 percent minimum O-ring squeeze. Therefore the 7.54 percent has been established as the minimum acceptable requirement for both case and nozzle O-ring joints and verified by subscale testing and full scale experience.

¶On a 0.280 inch diameter O-ring a 7.54 percent squeeze would be equal to a compression distance of 0.021 inches.22

¶On July 17, 1985, Irv Davids, Manager of the Solid Rocket Boost- er Program at NASA Headquarters, sent a memo to the Associate Administrator for Space Flight, the subject of which was case-to-Thiokol, Philip Shadlesky, "Performance Characteristics of the SRM O-ring Assembly Test Plan". TWR-14336. dated March 1984. D. 1. ~ *O fiiokol S. Rc&ers, "Significant Problem Report DR4-5/35 5 Day Re rt O-ring Erosion at Nozzle/Ah Segment Joint of SRM 11A (STS 41-BIMiasion 4143, &-i4370-1, May 7, 1984, pp. 1-2.

¶a 1 O-ring squeeze is the distance, in fractions of an inch, that an O-ring is compressed from its normally round shape. This dimension can also be expressed as a percentage of the total diame- ter before compresslon. In 1984 NASA waa using a term "minimum O-rin squeeze." During an SRM design anal sis of the cast? and nozzle O-ring joints it was conclufed that the 146 inch diameter cast? cygnders would not meet the design standard of 15 percent mlnimum O-ring squeeze at zero r w u r e . The various problems that prevented this included flaws in the O-ring y v e a and se%ng surfaces and differences in the spacing between tang and c l e w on various ?%kA, Larry Mullo "ECPSRM 1197, Nozzle Nose Inlet Housing O-ring Squeeze," SA 42- 562-84, November 20, 19&, p. 1.

195

¶case and nozzle-to-case O-ring seal erosion pr0blems.~3Davids sent copies to Messrs. Weeks, Hamby, Herrington and Winterhalter.23" In the memo it was noted that there has been twelve instances of primary O-ring erosion during Shuttle flights. In addition, in one specific case there had also been erosion of the secondary O-ring seal. There were also two primary O-ring seals that were heat af- fected without erosion and two cases in which soot blewby the pri- mary seals. In this memo it was noted that the prime suspect for the cause of erosion on the primary O-ring seals was the type of putty being used. It was Thiokol's position that during assembly leak check, or ignition, a hole could be formed through the putty which then initiated O-ring erosion due to a "jetting effect." It was even mentioned in this memo that Thiokol was seriously consider- ing the deletion of putty on the QM-5 nozzle/case joint since they believed the putty was the prime cause of the erosion. Davids, how- ever, had reservations about deleting the putty because he recog- nized the significance of the QM-5 firing in qualifying the FWC (Filament Wound Case) for flight.

¶In the matter of case-to-case O-ring erosion the memo noted that there had been five occurrences during flight where there was pri- mary field joint O-ring erosion. There was also one case where the secondary O-ring was heat damaged with no erosion. The memo stated:

The erosion with the field joint primary O-ring is consid- ered by some to be more critical than the nozzle joint due to the fact that during the pressure build up on the pri- mary O-ring the unpressurized field joint secondary seal unseats due to joint rotation.24

¶The memo continued:

The problem with the unseating of the secondary O-ring during joint rotation has been known for quite some time. In order to eliminate this problem on the FWC field joints a capture feature was designed which prevents the second- ary seal from lifting

¶Lastly the memo noted:

The present consensus is that if the primary O-ring seats during ignition, and subsequently fails, the unseated sec- ondary O-ring will not serve its intended purpose as a re- dundant seal. However, redundancy does exist during the ignition cycle, which is the most critical time.2s (See A p pendices VII-B and VII-C.)

¶On August 2, 1985, Larry Wear, MSFC's SRM Element Manager, sent a letter to Joseph Kilminster, Thiokol's Vice President for Space Booster Programs, on the subject of SRM field joint second23 NASA, Irving Davids, "Case to Case and Nozzle to Case '0Ring Seal Erosion Problems," July 11 1985.

23. Mr. Weeks, Dep. Assoc. Administrator for Space Flight (Technical); Mr. Hamby, Dep. Dir., STS Program Integration; Mr.Herrington, De Dir. of Launch & Landing Operations; and Mr. k. Winterhalter, Acting E r . , Shuttle Propulsion z 4 hid., p. 2.⁸⁶⁸⁶ Ibid., p. G I . Ibid., Volume I, p. 101. bid.

¶hid.

196

¶ary O-ring lift-off during pressurization. The letter concerned the situation wherein one O-ring might not seal subsequent to joint ro- tation. The letter stated:

Because of recent experiences of flight and ground test motors having increasing incidences of putty blow-holes and the associated burning of primary O-ring, it would seem prudent for us to attempt to assure that the second- ary O-ring is capable of sealing during the entire SRM

¶The letter requested an assessment of the possibility of lift-off of the secondary O-ring.

¶In August of 1985 Jim Thomas, MSFC's Deputy SRM Element Manager, wrote a memo for Mr. Mulloy to Mr. Hamby at NASA Headquarters, which was apparently never signed or sent. The sub- ject of the memo was SRM Joint/O-ring Erosion. The memo stated

On July 11, 1985, you and Irv Davids were briefed by Jim Thomas of my office on the history of the effort under- way to resolve the issues and concerns of the above sub- ject.

¶The memo than went on to discuss a number of questions.

1. What would happen if the secondary seal lifted off the mating surface during motor pressurization, and, also, how long it would take for the seal to return to a position where contact was made? The answer to that question stated that bench test data indicated that the O-ring resiliency, that is, its capability to fill the gap be- tween the tang and the clevis, was a function of temperature and the rate at which the gap opened.

¶The memo stated, "at 100 deg. F the O-ring maintained contact. At 75 deg. F the O-ring lost contact for 2.4 seconds. At 50 deg. F the O-ring did not reestablish contact in 10 minutes at which time the test was terminated." The memo then stated, "the conclusion is that secondary sealing capability in the SRM field joint cannot be guaranteed."²² Rogers Commission Report, Volume I, p. 199. 8

2. Another question concerned whether or not the secondary 0- ring would seal in sufficient time to prevent joint leakage if the primary O-ring had not sealed. The answer to that question was as follows:

MTI has no reason to suspect that the primary seal would ever fail after pressure equilibrium is reached, i.e., after the ignition transient. If the primary O-ring were to fail from 0 to 170 milliseconds, there is a very high proba- bility that the secondary O-ring would hold pressure since the case has not expanded appreciable at this point. If the primary seal were to fail from 170 to 330 milliseconds, the probability of the secondary seal holding is reduced. From 330 to 600 milliseconds the chance of the secondary seal

¶*' NASA, Larry Wear, "SRM Field Joint Secondary O-ring Lift-Of€ During Pressurization," SA 41-326-85,. August 2, 1985.²² Rogers Commission Report, Volume I, p. 199. 8 Engmeering consultants to the Committee have serious questions 88 to how this test relates to actual O-ring performance in flight hardware.

197

holding is small. This is direct result of the O-ring's slow response compared to the metal-case segments as the joint rotates.

3. The third question indicated that NASA Headquarters was not aware that the secondary O-ring may not seat due to joint rotation, and they wanted to know when this data was incorporated into the FMEAKIL? The answer noted that Thiokol had submitted a TWR- 13520 to MSFC in December of 1982. This was approved by NASA Level I11 on January 21, 1983. NASA Level I1 authorized a change request March 2, 1983 and Level I1 issued a PRCBD to implement approved Level I change request on May 2, 1983.29

¶Thiokol completed their engineering study of O-ring compression set and dated the report October 2, 1985.30 (Compression set relates to the ability of a material, in this case, O-rings, to rebound to its original dimensions after having being subjected to compression for various periods of time and or at various temperatures.) That report contained the following information. There was a concern of the ability of the O-ring to rebound to or near its original dimen- sions after having been subjected to compression for various peri- ods of time and at various temperatures. The Parker Seal Company of Culver City, California, tested several O-rings to determine the properties of the material. Two compression set tests in accordance with ASTM (American Society for Testing and Materials) D-395 method B were performed. The first test was conducted at a con- stant temperature of 75 deg. and the time that the ring was in compression was varied. In the second test the temperature was varied and the compression was held constant. A small O-ring of 0.139 inch diameter was used for test purposes. The test showed that the percentage of compression increased with an increase of temperature. However, these tests were not conducted at low tem- peratures. Rather, they were conducted at temperatures of 212 deg. F and above and therefore, they have little relevance to ambient conditions.

¶A status report from Thiokol's SRM O-ring Task Force, present- ed on November 20, 1985, recommended that a slightly larger Viton O-ring of 0.292 inch diameter, along with thicker shims, be used as a short-term solution. The current O-rings were 0.280 inches. Thiokol pointed out that there would be more erosion margin due to greater material thickness at the sealing surface. They noted that the thicker shims would reduce the initial and ab- solute final gap opening dimension, resulting in more O-ring "squeeze" initially. Thiokol stated that the greater initial squeeze would be better for compression set and resiliency, and would give a higher probability of maintaining a secondary seal longer into the ignition transient. Thiokol also noted that various tests were conducted on the Randolph putty using hot five-inch char motors. Two tests were conducted, which determined that the en NASA, Larry Mulloy, "SRM JointDring Erosion," SA 42-349-85, Au st 1985, pp. 1-2.

¶$ 0 Thiokol, B.L. Orme, "Enginering Study of O-ring Compression Set," &R-15218, October 2, 1985.³³ NASA, "Report to the President Actions to Implement the Recommendations of the Presi- dential Commission on the Space Shuttle Challenger Accident," July 14, 1986. (Hereafter r e ferred to as NASA Response to Rogers Commission.) * Refer to Appendix VII-A for ASTM specification.

¶JzSmall scale teat motors.

198

¶putty erosion could take place a t a rate between 5.5 and 13.0 mils per second. Two other tests noted that the erosion on GS-43 33 putty was ten times higher than that on the Randolph.34

¶Primary concerns drawn from the charts provided by Thiokol on January 27, 1986, centered around the following items. During the ignition transient, 0 to 170 milliseconds, there is a high probability of a reliable secondary seal. Between 170 and 330 milliseconds there is a reduced probability of a reliable secondary seal and be- tween 330 and 600 milliseconds there is a high probability of no secondary seal capability. Under steady state conditions, between 600 milliseconds and two minutes, the notes states "if erosion pene- trates primary O-ring seal-high probability of no secondary seal capability."³⁵³⁵ Most military and civilian aircraft engines are designed from the bottom-up approach, in which each component, starting with the material used all the wa through engineering testing of subsystems and subcomponents, is evaluated prior to the finardesign of the entire engine.

¶A. Bench testing showed O-ring not capable of maintaining con- tact with metal parts gap opening rate to MEOP.

¶B. Bench testing showed capability to maintain O-ring contact during initial phase (0 to 170 ms) of t r a n ~ i e n t . ~ ~

¶What follows is taken from Chart 2-2:

1. A temperature lower than current data base results in changing primary O-ring sealing timing function.

2. SRM 15-A 80 deg. arc black grease between O-rings. SRM 15-B 110 deg. arc black grease between O-rings.

3. Lower O-ring squeeze due to lower temperature. 4. Higher O-ring Shore hardness. 5. Thicker grease viscosity. 6. Higher O-ring pressure activation time. 7. Activation time increases, threshold of secondary seal pressurization capability is approached. 8. If threshold is reached then secondary seal may not be ca- pable of being pressurized.

¶The presentation went on to included the following blow-by history:

SRM 15 worst blow-by. A. Two case joints (80 deg.), (110 deg.) arc. B. Much worse visually than SRM 22. SRM b l o w - b ~ . ~ ~

¶The presentation then included a chart titled "O-ring (Viton) Shore Hardness vs. Temperature." 39

Degree F Shore Hardness 70 degrees 77 hardness 60 degrees 81 hardness 50 degrees 84 hardness 40 degrees 88 hardness 30 degrees 92 hardness 20 degrees 94 hardness 10 degrees 96 hardness

¶33 A type of, utty made b another company that also was considered for use in the SRM.³⁴³⁴ Ibid., pp. K-23 through K-27; K-31. Thiokol, 8RM O-ring $ask Force Status and QM-5 Recommendations," TWR-15349, No- vember 20, 1985.

¶Thiokol. "Temuerature Concern on SRM Joints," January 27, 1986, chart 2-1.³⁵³⁵ Most military and civilian aircraft engines are designed from the bottom-up approach, in which each component, starting with the material used all the wa through engineering testing of subsystems and subcomponents, is evaluated prior to the finardesign of the entire engine. -..~

¶~

¶30 bid.³³ NASA, "Report to the President Actions to Implement the Recommendations of the Presi- dential Commission on the Space Shuttle Challenger Accident," July 14, 1986. (Hereafter r e ferred to as NASA Response to Rogers Commission.) 7 bid., Chart 2 2.

¶3s bid.. Chart 3-1.

¶39 Kid.; Chart 4-1

199

¶The term Shore Hardness refers to a method of identifying the hardness of materials, and a higher number means a harder mate- rial. Regardless, though, it is seen from the above table that the hardness increases as the temperature decreases. Ecgineers pre- sented a chart titled Secondary O-ring Resiliency, listing the fol- lowing temperatures. O

Temperature degree F Time to recover (seconds) 50 degree 600 recover 75 degree 2.4 recover 100 degree *did not separate

¶The conclusions presented at the end of the teleconference were:

1. Temperature of O-ring is not only parameter controlling blow-by. SRM 15 with blow-by at an O-ring temperature at 53 deg. F. SRM 22 with blow-by at an O-ring temperature at 75 deg. F. Four development motors with no blow-by were tested at O-ring temperature of 47 deg. to 52 deg. F. Development motors had putty packing which resulted in better perform- ance. 2. At about 50 deg. F blow-by could be experienced in case joints. 3. Temperature for SRhl 25 on 1/28/86 will be 29 deg. E' 9:OO a.m.,³²³² Bid., p. 291. deg. F. 2:OO p.m. 4. Have no data that would indicate SRM 25 is different than SRM 15 other than t e m p e r a t ~ r e . ~ ~

¶Recommendations

1. O-ring temperature must be greater than or equal to 53 deg. F at launch. Development motors at 47 deg. to 52 deg. F with putty packing had no blow-by. SRM 15 (the best simula- tion) worked at 53 deg. F. 2. Project ambient conditons (temperature and wind! to de- termine launch time.42

¶The effect of Thiokol's recommendations would be that the Shut- tle should not be launched unless the O-ring seal temperature was at least 53°F. (e) Loads Acting on the Joint

¶There are other loads on the joint in addition to those caused by the pressures of the burning propellant. The following table identi- fies those loads relative to time.43

¶Time Activitv Source 01 load Static or dvnamtc lmwct on ioint

¶Days before launch .......... Mating of casing.............. Weight of upper casing Static plus impact ........ Physical contact

contacting lower between tang and casing. clevis.

¶40Ibid., Chart 4-2.

¶41Ibid., Chart "Conclusions."⁴⁴ Rogers Commission Report, Volume 11, p. H-1. 2 Ibid.. Chart "Recommendations."⁴⁴ Rogers Commission Report, Volume 11, p. H-1. 3 This'Chart was prepared by the Committee and is based on information obtained by Com. mittee staff during meetings at MSFC on June 30, 1986.

200

¶Time Activity Source of load Static or dynamic Impact on joint

¶Days before launch .......... Solid rocket motor Weight of SRB Static ............................ Compressive: shear on

assembly (Stacking) components. pins at faces at KSC. between tang and each clevis leg.

¶Days before launch .......... Joining of the external Additional weight of tank Static ............................ Compressive; additional

tank, and orbiter to and orbiter. shear on pins which SRBs. connect tang to clevis.

¶Days before launch .......... Transport to pad on Movement of Transporter ... Static and dynamic....... Compressive; slight

crawler. shear changes on pins.

¶Days befcre launch .......... Addition of payloads ........ Added weight ..................... Static .......................... Additional compressive and shear loads. Within 24 hours of Loading of fuel ................Weight of liquid hydrogen Static.......................... Additional compressive launch. and liquid oxygen. and shear loads. Within 24 hours of Loading of fuel ................ External tank contracts in Static ............................ Lateral tensile force launch. diameter due to applied by aft

reduction in attachment structure temperature. between external tank and solid rocket motor casing.

¶6 semds to launch ........ Firing of main engines Thrust of engines ............... Static and dynamic....... Further moments

(SSME's). compressive and vibratory (25 to 30 Hz) .

¶At start of launch............ Solid rocket motor Combustion pressures......... Basically static ............ Bending (Joint rotation)

ignition before lift-off. lateral forces perpendicular to casing walls.

¶At start of launch .......... Solid rocket motor Engine thrust ..................... Static and dynamic....... Instant load reversal

ignition before liftoff. from compressive to tensile in joints and instant shear reversal in pins.

¶At start of launch............ Release of hold down Instant release of lateral Dynamic........................ Instant change in

bolts. force at aft External stress, vibratory at 3 Tank attachment Hz. structure.

¶At liftoff ........................... Launch maneuvering........ Thrust, plus nozzle vector Static and dynamic .......

forces. Combination: tensile, shear, Vibration, lateral (via attachment structure).

¶Launch phase...................In-flight maneuvering....... Thrust plus gimbaling, Static and dynamic....... Cornbination vibration,

applied loads at tensile, shear, lateral. attachments.

¶Launch phase................... Turbulence-wind gust Impact, thrust nozzle Static and dynamic....... Impact loads

loads. gimbaling (changes in transmitted to joints. applied loads). .... Reduction of main Decrease in thrust .............. Static and dynamic....... Changes in bending and engine wwer and stress in joint, Sogd Racket Motor changes in frequency thrust at Max q of vibration. (maximum dynamic pressure).

¶Launch phase................... Increases in thrust Increase in thrust .............. Static and dynamic....... Changes in bending and

thrust of main stress in joint and in engines and solid vibration. rocket motors.

¶Separation phase.............. Burning out of solid Relsase of thrust, impact Static and dynamic....... Reduction of tensile

propellants and forces at attachment forces and shear on explosive forces at points. pins. attachment points.

20 1 - Time Activity Source of load Static or dvoamic lmmt rn mnl

¶Ocsan impact and Weight of SRB Impact on joint ................. Dynamic. ...................... Variation in stresses at retrieval. impacting Ocean at joints

about 60 mph (vertical) 25 mph (horizontal).

¶How these loads are accommodated by the joint is critical to the seal. In Thiokol's analytical evaluation report (TWR-12019, dated October 6, 19781, S. Stein of the Structures Section included the statement, "except in local area of pin, all stress levels are consid- erably below yield." 43a As a result of this information, the Com- mittee will explore this condition as part of its normal oversight work to determine the long-term effect on structural integrity of the casings.44 Stein also wrote, "at MEOP [maximum expected op- erating pressure] the primary '0'ring gap increases 0.052 and the secondary 0.038"."⁴⁴ Rogers Commission Report, Volume 11, p. H-1. 5 It should be noted however, the analysis was made for no thrust, i.e. internal pressure As noted on the forgoing chart, loads on the joint do work in combination and so the analysis should also provide for the combined effect of all loads at the time they occur.

¶On page 55 of the Rogers Commission Report there is a chart which shows a series of curves which relate maximum aerodynam- ic force to Mach Number. As a result of a discussion with Dr. Rich- ard Feynman, Department of Physics, California Institute of Tech- nology, and a member of the Rogers Commission, the Committee will review these curves after the completion of this report in an effort to ascertain their validity. There is reason to suspect that the "flight envelope'' as repressented in the chart is ina~curate.~'

¶As stated previously, the proper choice of materials is critical to attaining performance objectives. The steel casings are designed to withstand the propellant pressures and loads incurred in flight. Secondly, they must accommodate these forces over and over as the casings are reused. Consequently, the choices of the type of steel selected was important.

¶The steel used to make the casings and the joint is a D-6A. D-6A is a low-alloy steel for aircraft and missile structural applications. It is designed primarily for use at room-temperature tensile strengths of 260 to 290 k ~ i . D-6A

¶~ ~maintains a very high ratio of yield structure to tensile strength up to a tensile strength of 280 ksi, combined with good ductility.

¶Typical mechanical properties of D-6A steel:⁴⁴ Rogers Commission Report, Volume 11, p. H-1. 9

¶"Thiokol, S. Stein, "Analytical Evaluation of the Space Shuttle Solid Rocket Motor Tang/ Clevis Joint Behavior", TWR-12019, October 6, 1978, p. 1.⁴⁴ Rogers Commission Report, Volume 11, p. H-1. 4 Thiokol, S. Stein, "Analytical Evaluation of Space Shuttle SRM Tang/Clevis Joint Behav- ior", TWR-12019, October 6, 1978.

¶'5 Ibid.⁴⁴ Rogers Commission Report, Volume 11, p. H-1. e Ibid.⁴⁴ Rogers Commission Report, Volume 11, p. H-1. 7 Discussion with Dr. Richard Feynman, California Institute of Technology, Pasadena, Cali- fornia, September 3, 1986.⁴⁴ Rogers Commission Report, Volume 11, p. H-1. 8 1000 pounds per square inch equals 1 ksi.⁴⁴ Rogers Commission Report, Volume 11, p. H-1. 9 American Society for Metals, Handbook Edited by H.E. Boyer and T.L. Gall, November, 1984.

202

Temoerinc temcerature Tensile strength yield strength ks, E'a2i~n,50 Reduction in area, V=nolch impaci 'C 7 ksi cercenl percent energy fl-lb

¶

• 150 300 299 211 8.5 19.c — 10

• 205 400 290 235 8.9 25.7 — 11

• 315 600 267 24 7 8.1 30 0 — 12

• 425 800 236 228 9.6 36.8 — 12

• 540 1,000 210 204 13.0 45.5 — 19

• 650 1,200 150 141 18.4 60.8 — 30

¶Normalized at 910°C (1650'F) and tempered at various temperatures

¶In addition to the steel, other principal materials in the joint design that were to seal in the propellant gases were the zinc chro- mate putty and the O-rings.

¶On April 12, 1984, John Miller, Chief of the Solid Motor Branch of NASA, wrote a memo to Mr. Horton, Chief Engineer, SRB Engi- neering Office, MSFC which referred to concerns with putty made by Randolph. The Randolph putty was selected on the basis that it had several desirable performance characteristics. The change in putty was made after Fuller-O'Brien discontinued making putty be- cause their product contained asbestos. Mr. Miller noted, "Stacking difficulties and observed O-ring anomalies appear to be more fre- quent with Randolph putty than with the previously used Fuller- O'Brien putty."⁵⁵ bid. 0 %id. iNm.-The nozzle to case joint design is significantly different than the case field joint design w ich caused the Challenger accident. However, it is cited here because some of the prob- l e m are relevant to the failure of the aft field joint.] O Miller requested that Thiokol expedite develop- ment and qualification of a putty with properties similar to those of Fuller-O'Brien.

¶On June 18, 1984, Miller wrote Horton again, mentioning ero- siodheat exposure O-ring experience on QM-4, STS-2, STS-6, STS-11, and STS-13 and citing Deficiency Reports which violated specification^.^

¶BY June 29. 1984,⁵⁵ bid. 0 %id. iNm.-The nozzle to case joint design is significantly different than the case field joint design w ich caused the Challenger accident. However, it is cited here because some of the prob- l e m are relevant to the failure of the aft field joint.] inch motor tests has been completed. These tes& substantiated the concept of hot gas jet impingement against O-rings. Interestingly, a simulation of "no putty' yielded no O-ring damage. This information was conveyed to NASA via telecon from Thiokol, which also stated that there was no second source for the Randolph putty. Thiokol had abandoned their program to mix the putty themselves. Measures taken to correct the putty problems in- cluded changes in the putty layup to reduce air entrapment, use of a porous sacrificial heat barrier such as carborundum fiberfrax or removing the putty and reducing joint gaps were introduced.5 2

¶A new joint design was forwarded to NASA by Thiokol on July 19, 1984, which included a fill capture feature. This feature looked similar to the "capture feature" proposed for future Shuttle flights. The fill capture feature, however, was to be filled with grease. A thermal analysis had shown that "severe heat effects would result if the cavity were not filled."

¶As stated previously, the putty was to insulate the O-ring seals from the hot propellant gases. It was also to remain flexible enough to move outward under the pressure of the burning propel- lant, thereby compressing the gas in the joint which, in turn, was

¶5 0 NASA, John Miller, "Concerns with Randolph Vacuum Putty," EP-25 (84-35), April 12, 1984.⁵⁵ bid. 0 %id. iNm.-The nozzle to case joint design is significantly different than the case field joint design w ich caused the Challenger accident. However, it is cited here because some of the prob- l e m are relevant to the failure of the aft field joint.] 1 NASA, John Miller, "Zinc Chromate Putty Installation in Nozzle to Case Joint Discrepan- cy," EP-25 (84-53), June 18, 1984.⁵⁵ bid. 0 %id. iNm.-The nozzle to case joint design is significantly different than the case field joint design w ich caused the Challenger accident. However, it is cited here because some of the prob- l e m are relevant to the failure of the aft field joint.] 2 Thiokol, "Vacuum Putty Telecon," June 29, 1984.

203

¶to seat the primary O-ring. O-rings require some pressure from the working fluid, in the case of the SRM, this was gas, in order to seat properly and provide a effective seal. In practice, this design philos- ophy did not prove to be correct because the putty frequently held the pressure off of the O-rings, or if it did not, the putty had blow- holes in it. It was then postulated that these holes might actually benefit the seating of the O-ring by allowing more pressure to reach it sooner. It was even suggested that holes might be deliber- ately made through the putty. However, it was then learned that blowholes served to concentrate propellant gas on small segments of the primary O-ring and caused the ring to erode.

¶The unacceptable heat erosion damage to both primary and sec- ondary O-rings on SRM-16A resulted in an evaluation of the putty produced by Randolph Products. In July 1985, L. Thompson of MSFC made a presentation which noted that five different types of putty from four companies were under study in an effort to solve the putty performance problem. As late as 1985 twelve different types of tests had been performed and six more were in progress. The only putty to survive the water tests was General Sealants No. 43, which was a non-asbestos formulation. The Randolph putty had disintegrated in all three water tests. However, in comparing dy- namic viscosity to temperature, the General Sealants product, at 25,000 poise,53 was not viscous above 125 deg C. It was slightly better than the Randolph product and another product made by Inmont. The previously used Fuller-O'Brien product, however, in- creased in dynamic viscosity with an increase in temperature. It was 100,000 poise at 250 deg C, while it was less than 50,000 at 50 deg C.54Consequently, no product met all the design requirements as well as the Fuller-O'Brien did.

¶The Randolph putty is hydroscopic and its behavior is unsuited to use in the dry climate of Utah, as well as the humid climate of the Florida coast. In one case the putty was too stiff and in the other, too sticky. Since both factory and field joints required the use of the putty, a product with consistent performance in both cli- mates was required.

¶The materials used in the manufacture of the O-rings was also critical to the safe operation of the Shuttle system. The O-rings had to be serviceable at the high temperatures in the joint which would result from heat transfer from the rocket combustion chamber. However, the use of NBR insulation around the propellant, and the use of putty, was to protect the steel casings and the O-rings from the direct heat of the propellant gases. This protection was not always successful when blowholes in the putty occurred, however, and the O-rings would frequently be damaged by heat. The lower temperatures that occur in Florida during the winter months was not covered by NASA's specifications. While elastomers are known to become brittle at low temperatures, a product specification sheet on Viton Fluroelastomer claimed, "Cold-VITON is generally serv- iceable in dynamic applications down to - 18 to -23 deg C (0 to 10 deg F)."55The sheet added: "The brittle point of Viton at a thickness of 0.075 inches is in the neighborhood of 50 deg F. Yet, as with other elastomers, thickness has a marked effect upon low tempera- ture flexibility. Thinner cross-sections are more flexible than thick- er ones at every temperature." The thickness of the O-rings on the Shuttle is 0.280 inches, thicker than the 0.075 inch article with a brittle point of 50 deg F noted above.56 Consequently, the brittle point of Viton was misleading since the O-rings were much larger the the test specimen.

204

¶Military Specification MIL-R-83248A, 17 Feb. 84, "Rubber, Fluo- rocarbon Elastomer, High Temperature, Fluid, and Compression Set Resistant" set the specification for the O-rings that Thiokol had to meet.57 They included:

Type I-O-rings and compression seals Class 1-75 +/ -5 Hard- nessS8

¶This specification then included other specifications issued by the Society of Automotive Engineers and the American Society for Testing Materials. One of the ASTM Specifications listed was ASTM 1329, "Evaluating Rubber Property, Retraction at Low Tem- peratures."S9 It was these referenced specifications which defined the significant characteristics required.

¶On February 6, 1979, Mr. William Ray of NASA's Marshall Space Flight Center, sent a memo to Messrs Hardy, Rice, Eudy, and McCool (See Appendix V-I). That memo was essentially a trip report of Mr. Ray's visits to the Precision Rubber Products Compa- ny and the Parker Seal Company, in search of information on the performance of O-rings. Some of the points covered in the memo were:

The purpose of the visits was to present the O-ring seal manufacturers with data concerning the large O-ring ex-

¶trusion gaps being experienced on the Space Shuttle Solid

¶Rocket Motor clevis joints and to seek opinions regarding the potential risks involved.60

¶With regard to the visit with company officials at Precision Rubber Products, "they voiced concern for the design, stating that the SRM O-ring extrusion gap was larger than

that covered by their experience." In response to the data presented to Parker Seal Compa- ny officials by Mr. Ray, Parker officials "also expressed

¶surmise that the seal had wrformed so well in the present

appiication." 6 2 Regarding the visit with the Parker officials, the memo

¶stated. "their first thought was the O-ring was being asked

to perkorm beyond its intended design and that a different type of seal should be considered." g 3

205

¶The need for additional testing of the present design was also discussed and it was agreed that tests which more closely simulate actual conditions should be done.s4

¶As a result of the foregoing data, the Committee has arrived at the specific Findings and Recommendations contained in Chapter V.

LAUNCH OPERATIONS * INTRODUCTION

¶The purpose of this section is to document the series of decisions that culminated in the launch of STS 51-L on January 28, 1986. In Section A, the discussion details the Flight Readiness Reviews used to assess the mission's readiness, and also describes the teleconfer- ence on the night of January 27 when Thiokol engineers attempted to delay the launch. Also discussed are the circumstances sur- rounding the uncertainty represented by ice covering the launch pad's gantry.

¶Section B describes a specific example where the launch crew in the Firing Room waived a launch commit criterion. The discussions that took place on the subject indicate that the alternate procedure used as a justification for the waiver should not have been allowed, since the environmental conditions on the morning of January 28 were outside the limits specified for the alternate procedure.