[M1] Morton Thiokol was invited by the Commission to review and comment on the NASA Accident Analysis Team Report and the NASA Solid Rocket Motor Working Group Report. Mr. Arnold Thompson reviewed the Accident Analysis Team Report on April 24, 1986. Messrs. Edward Dorsey, Roger Boisjoly and Allan McDonald reviewed and made notes on both reports on May 8, 1986. Messrs. Edward Dorsey, Thompson, McDonald, Boisjoly, and Ed Garrison (all of Morton Thiokol) then discussed these notes with the Commission on May 9, 1986.

The Commission agreed to include these comments to assure a complete historical record. The original page and paragraph numbers noted by the author for each comment correspond to the basic report originally submitted by the STS 51-L Data and Design Analysis Task Force Accident Analysis Team, including the separate report of its Solid Rocket Motor Working Group. These reports have been reformatted and reproduced in their entirety as Appendix L of this volume. The parenthetical inserts, such as "(L-65, para. 3)," indicate the corresponding location of the cited material in relation to the format of Appendix L.

[M3]

EXECUTIVE SUMMARY

SYSTEMS REPORT (Not reproduced in Commission Report)

1. Even after corrections are made, RH aft field joint is the coldest.

2. As far as negative mismatch is concerned there were 7 or 8 joints in previous flights worse than STS-51L (STS-9 was-0.540" for left aft center). Sketch needs to show chamfer on inner clevis leg-misleading without this!

3. Ice team document (p.56 Table 3) Vol. II indicates possible correlation between large negative IR temps and ambient for 51C, 51B and 51L. 51B raises questions about need for corrections to IR data.

4. Did anyone ever check if 16°F antifreeze isn't really better than that, i.e. maybe 8-10°F (mfg. usually has some margin) would be good reference.

5. How do water troughs radiate to sky-aren't they basically under the nozzles and Shuttle vehicle? Why didn't south troughs freeze? Where are they?

6. Why does ice team say ice is greater than 55 lb/ft³ and MSFC uses 25 lb/ft³ for calculations?

7. What does (Qa + QB) for STS-51L look like relative to all previous flights? [a = alpha; B = beta]

8. Why does Shuttle fly beyond design limits for ET/SRB design envelope?

9. It's a fact that STS-51L was the coldest, had highest Qa highest QB, highest TVC gimbal profile-all of which most likely contributed to joint seal failure in RH SRB aft field joint-Why don't we say so?

10. Why were strut loads at liftoff so much higher on STS-51L compared to STS-1-7 and especially STS-6 w/o prelaunch loads and same payload?

11. p.61 Exec. Summary statement needs changing to most significant factor in temperature. (L-45; "Joint Temperature")

12. Statement on not realizing putty can seal before STS-51L failure is not true and should be changed.

[M9] 13. After temperature corrections for suspect "ice" measurements on LH SRB Temperature, correction came out to be 36°F not 32°F.

14. Test showing 4.1% stretch in O-ring and conditions for developing metal slivers from joint mismatch are not reasonable-restrained sections do not represent unrestrained full circumference joint mating.

Executive Summary

Paragraph 3.2 - Report gives 36°F + 15°F = 51°F for next coldest temperature at launch time STS-61C). MTI data shows-61C ambient temperature at launch as 55°F. (L-4)

Paragraph 4.0 - Last Sentence. A more descriptive wording might be as follows: "a failure in the right SRM aft field joint which resulted in a hot gas leak." (L-16)

Paragraph 4.3, Subparagraph a, Page 26 - Comment relative to the OMS: An Aviation Week article said there was a surprising amount of heat damage to the Orbiter tail (vertical stabilizer). Are the OMS pressure measurements such that propellant leaks could be detected during the ascent flight? (L-23)

Paragraph 5.2, Page 51 - ".........A flaw 0.1 inch long and 0.50 inch deep.......should read,..........and 0.050 inch deep..........." (L-35)

Paragraph 6.0, Page 57, 3rd Paragraph - Statement "using specific STS-51L hardware configuration" implies that the extensive tests each had the exact -51L hardware configuration. I don't believe that to be the case. (L-37)

Paragraph 6.1, Page 58, 1st Paragraph - This discussion is not as clear as it should be. For example, I think the "flat-on-flat" prohibition would prevent joint assembly if the diameter negative differences were as much as quoted here. Also, Figure 6.1.1 does not show the chamfer which exists on the female clevis inner leg. (L-37)

Paragraph 6.1. Page 58, 2nd Paragraph - The pinch marks on SRM 14A could have resulted from O-ring extrusion into the gap and subsequent "nibbling" as the motor pressure decayed and not necessarily from assembly damage. In the ground test program, I only remember one instance of O-ring damage, and I think it was attributed to disassembly, not assembly. (L-42)

Paragraph 6.1, Page 58, 3rd Paragraph - An O-ring length increase of 4.1% in a full diameter segment would literally result in the O-ring falling out of the groove over a significant distance. Nothing of this magnitude has ever occurred during an assembly process. (L-42)

[M11] Accident Analysis Team Report

04-24-86

Page 2

Page 59,1st Paragraph - The possibility also exists that 0.001 to 0.003 inch slivers will not cause a leak under dynamic conditions. The dynamic test data should be added to the report if available or we should not speculate on the results (L-42; 3rd full para.; left column)

Paragraph 6.3, Page 60, Last Paragraph, 4th Line - Should read ".......upstream face of the O-ring groove and the .........." (L-45; 1st para.;left column)

Paragraph 6.4, Page 61, 1st Paragraph, 5th Line - This sentence would be clearer if it read "...........predicted local temperatures of 28 ± 5°F at the coldest circumferential location." (L-45)

I have some additional concerns about the temperature discussions:

a. The sentence quoted above is still not clear as to whether it applies only to the right hand and left hand aft field joints or to all field joints.

b. The discussion is very abbreviated relative to the importance of possible local temperature effects. For example, nothing is said about possible local cold pockets caused by the venting of cryogenic vapors or air flow past the cold ET. Nothing is said about the freezing of the water/anti freeze mixture in the sound suppression troughs.

c. A Morton Thiokol assessment is that the right hand aft field joint could be in the range of 21 ± 5°F.

General Comment

Findings and conclusions in the Executive Summary were not available for review.

Original hand-written notes were signed by E.G. Dorsey

[M12] 04-25-86

HOSC/MSFC

J. Kilminster, E. Dorsey

Page 12, 1st Paragraph (L-56)

Comment: The times (MET) should all be cross-checked with the Executive Summary Report and the official MET for consistency.

Section 3. Page 22, 3rd Paragraph (L-62; 4th full para.; left column)

The description of O-ring damage is misleading. The pinch marks on SRM 14A could be the result of O-ring extrusion into the gap and subsequent pinching as the motor pressure decayed; not assembly damage. I think there was one instance of O-ring damage during ground test (static firing), and that happened during disassembly.

Page 24, 2-M Paragraph (L-62; 4th full para.; right column)

I would think the flat-on-flat prohibition applies to any flat-on-flat situation, including those caused by negative dimensions. It should also be noted that the tang end has chamfers to prevent O-ring damage during assembly.

Page 26, Last Paragraph (L-65; 4th full para.; right column)

My understanding is that metal slivers were generated under tang entry angle conditions which would not exist under actual complete segment assembly conditions. The reported 4.1% O-ring length increase in the short test section doesn't extrapolate to a reasonable value for a full diameter segment; i.e., the 4.1% applied over the full circumference would cause the O-ring to fall out of' the groove for quite some length. Nothing like that has happened.

Section 4d, Page 55, Item (3) (L-73; last para.; right column)

The launch temperature analysis reported here is a summary of the analysis in Appendix B and is too condensed relative to the importance of this subject. A table showing vehicle temperatures by location for both SRB's for 7:00 am and for launch time is needed, as well as a plot of circumferential temperature variation around the field joints. Also, nothing is said about the possibility of local cold pockets caused by the venting of cryogenic vapors or air flow past the cold ET. The freezing of the water/anti freeze mixture in the sound suppression troughs is not mentioned. In summary, there are reasons why the joint temperature could be less than 28° ± 5°, but a reader would not understand that from this paragraph. A Morton Thiokol assessment is that the right hand field joint could be in the range of 21° ± 5°

[M13] SRM Working Group Review

04-25-86

Page 2

Section V. Summary, Pages 77 through 82 (L-85 through L-88)

I have previously commented on each of the following potential failure cause or contributor, so these remarks will be brief:

A. Assembly Damage/Contamination - I agree that this is a potential failure cause and one that could be unique to a single field joint, but I don't agree that damage probability is as high as I can infer from this description. From past personal observation I don't believe the assembly crews would force fit a joint if they had a flat-on-flat condition resulting from negative diametrical measurements. Also, my memory is that there was one incident of O-ring damage (cutting) during ground static firing, and that happened when the joint was pulled apart. I also find it improbably that a full circumference O-ring would grow 4.1% during a joint assembly process.

B. Gap Opening - No comment.

C. O-Ring Squeeze As Mated (Static) - No comment.

D. Joint Temperature - As previously noted, there is a possibility of local cold pockets around an SRM because of cryogenic vapor venting and/or air flow past a cold ET. The right hand aft field joint could have been colder than 28° ± 5°, but that possibility is not discussed in the body of the report or in this summary.

E. Putty Performance - No comment.

Section VI, Findings (L-88 and L-89)

Following Finding #3, it should be noted that a joint leak occurred on the coldest launch of the Space Shuttle, and that there is a potential for local temperatures to be below the 28° ± 5° assessment for all joint temperatures.

Under Finding #4, an additional study should be made to determine if there is a correlation between previously observed instances of O-ring erosion and the location of the maximum interference or closest fit between the joint tang and inner clevis leg. The existence or absence of such correlation should be stated in the finding.

Original Hand-Written Notes were signed by E. G. Dorsey

Tests to simulate joint damage were done at 1-1/2°, the maximum allowed is 0.39° (1" across 146").

A flat on flat of 0.005" to 0.010" was forced on clevis leg which caused slivers.

The joint is assembled with a gentle wipe in action that does not raise metal burrs or slivers. The radial assembly forces are very low between tang and clevis Parts.

The slivers are formed when the pins are inserted. These slivers are below the seal zone.

Post flight inspection of SRM-14A field joint o-ring showed pinch marks. I believe these marks were caused by the o-ring extruding into the gap under pressure. When the pressure was released, the gap closed nibbling the o-ring material. Photos from testing show this phenomena.

The 4.1% increase in o-ring length would be 19". I think this is a gross over-statement.

The test data results of O.001 > 0.003 sliver contamination should be added to report.

Diameter difference causing non-concentric is not true. It expresses elliptical shape not concentricity.

Last paragraph at the bottom of page 60 - I believe that care should be exercised in this area so an to not misguide people about o-ring sealing mechanics. The timing requirement determined by resiliency tests is very critical. The time that the o-ring is pressurize Vs the gap opening sequence determines whether blow-by will occur. Some carefully designed tests and analysis are necessary to fully understand the sealing characteristics of the o-ring.

The write-up shows the MSFC resilience data which the rubber industry uses to compare materials. I feel that the work done at MTI better explains the o-ring's resiliency to define the time at which separation of the o-ring with its tang sealing surface occurs. Our data shows at about 200 m-sec the seal separates. It also shows an increase response with greater squeeze.

The issue of ice in the joint interfering with the secondary o-ring, I believe, is out of proportion. The joint would have to be tilted 3/8" under static load to allow this to happen. The posts without load are level within 0.005". Some bending deflection will increase this out-of-level condition.

First paragraph talks about 530 m-sec and 1.9 sec delay of pressure to o-ring. Compared with 200 m-sec seal opening both of these times will allow blow-by. I don't understand why 0.010" is particularly significant.

These conditions could cause initial blow-by. Once blow-by has initiated, no guarantee exists that the seal will be effective.. First, the hot blow-by gases could cause seal damage due to blow-by erosion., Second, the pressure forces on the o-ring, depending on the time that the seal sees Pressure and the amount of gap opening, it may never function properly.

May 6, 1986

Joint re-design meeting to present new design in preparation for the technical in the change meeting on 7 May 1986.

• Notably missing from this meeting was Al McDonald.
• I brought up both the gray tool company and Lokring design concepts that would prevent joint opening from occurring at all. Everyone in mtg is now aware that such a design possibility exists.
• Brian Russell called about the possibility of ordering some "ETPEL-F" material for our seal selection testing. I reviewed the material data package and said go ahead, it can't hurt because the properties are in the region that we are interested in.

May 7, 1986

• Travel day to Washington D.C.

May 8, 1986

• Review of reports submitted to Presidential Commission SRM-Accident Analysis Team Report.

Pg 22 (L-62; 3rd para.; left column)

• 2nd sentence "a review of the right..........failure of clevis leg" The mechanism is not based on out-of-roundness locations or magnitudes but upon not having a flat on flat condition so that the parts will mate by self aligning themselves, etc.
• Sentence "additional data were provided....to end of para- Some of data was in question and if one considers the affect of [M17] propellant wt causing the tang dia to increase and the inboard clevis leg dia to decrease, these would have a canceling effecting relative to potential interferences as stated.
• 3rd para - If this type of data is going to be used then a more full explanation needs to be given. Small metal slivers were found in the DM-5 joint disassembly but not across the seal is my recollection. (L-62, 4th para.; left column)

Pg 23 (L-62, last para.; left column)

• 2nd para - "The SRM case segments...........phase by 90°" - The shape of the segments is not the driver in mating - flat on flat is - our horizontal mates tart the TDC and then the remaining circ feed into place as the thin shell elements mutually deflect. The O-rings are not cut or damaged by this method.
• 3rd para - Flat on flat is the req'mt - the dimension is used only as a guide for mating.
• 3rd para - "however, no requirement...........Figure 10)."- Flat on flat applied to every surface in the mating plane.

Pg 24

• 1st para - "The 51-L right SRM...........excessive ovalaity" - I doubt this because it took days to assemble the joint for which the tool that is now used was designed for - I designed the tool and transmitted the info via phone to KSC myself. (L-62; 4th para.; right column)
• 2nd para - "Because of the extreme.......of the O-ring seals." The cases mate just fine as long as flat on flat does not exist. This is really reaching for mating causes creating problems. (L-62; Sth para.; right column)
• [M18] 3rd Para - "Although the..........after assembly" - None of these comes close to a horizontal mate condition. (L-62; 6th para.; right column)
• 4th para "Measurements including Pi tape...........at the clevis sealing surface" - I saw the data and first questioned way the dia changed from the restrained to the unrestrained condition and no one could give me an answer. I don't question the fact that dimensions are changing but that not enough time or thought has been given to explaining magnitudes and reasons. I believe the original ROHR diameters on the inbonded clevis leg are taken prior to machining the O-ring grooves - yet we do not know if just machining the grooves and softening, the hoop is responsible for some of the dia changes. (L-62; last para.; right column)

Pg 25 (L-65; 1st para.; left column)

• 1st para - "Extrapolation of the interference fit" - How can one extrapolate valves like these, except to make a predetermined point. The case segments are each unique as they are removed from the lathe - their free state action is not necessarily repeatable at all. The "rounding ring" is actually a "restraint ring" for machining purposes and as such does; not give the same free state end results. We do know that all circumferences are indeed good numbers.
• 2nd para - Same comment as 1st para (L-65; 2nd para.; left column)
• 3rd & 4th para - This data shows that very little force is req'd to move the joints to conform to one another: This also shows that slight interferences (should they exist) are not important to either loads or stresses. (L-65; 3rd & 4th pares.; left column)

[M19] Pg 26 (L-65; 4th full para.; right column)

4th para - "Two significant findings vulnerable to damage" this whole finding is a myth - 1st - by fixing two pieces of joint in a machine makes them so stiff that the), cannot mutually conform to one another which does not require much force as previously stated. Secondly - the elongation of the O-ring caused by this mechanism, if real as stated, would result in gross problems during all horizontal matings since the joints are actually mated by engaging at TDC and allowing the case to mate around the circ towards BDC at the end of closure.

Page 27 (L-65; 5th full para.; right column)

1st para - "The results of these tests ...........in the O-rings" It was my understanding that the short stack was forced to have metal to metal contact and forced to give the most change for the O-rings to exhibit damage but damage did not result except in special port locations that are not in an actual joint. I don't see why this data isn't applicableto show that the ass'y is very forgiving relative to damaging the seals.

Page 27 (L-65; 6th full para.; right column)

(5) 1 see the value of these tests but was sufficient time given after assly to allow the elastomer to flow around the metal sliver and then leak check to see if time has an affect on sealability.

Pg 28 (L-68; 4th full para.; left column)

1st para - "Further, the location...........interface at mating" óThis is a totally speculative statement with no basis since it has already been stated that very low forces are returned to make the joint conform during mating. This implies that the parts would move laterally before creating enough force to cut the metal at a specific location.

2nd para - Same comment

[M20] Pg 33 Questions on data table (L-66; bottom; Figure 12)

1. What was measurement basis of clevis gaps - was the segment vertical and supported 3600 and round within its mating dims? A .005" difference in clevis opening is easy to attain if the above was not done.

2. Were the original ROHR clevis and tang diameters measured prior to machining an O-ring groove and machining joint holes? If so, are the dims real or partly due to comparing apples and orange

3. What is the accuracy assessment of the measurements were measurements taken restrained or unrestrained, if restrained, was tool pressure the same as original measurements?

Pg 35 (L-68; 6th full para.; left column)

Why is scenario 4 limited only to primary O-ring blow-by since the puff of smoke indicates that the secondary O-ring was also by-passed. One branch obviously missing is temperature resiliency affects on both the primary and secondary O-ring.

3rd para - I don't understand the basis of these statements generally more squeeze better seal as long as overfill not created causing overstress in seal mat'l. This ties in low temp as a definite problem but may also be used to show that if the seal was touching both side walls then it would have been in the correct sealing position at ignition (no motion across groove req'd). This makes resiliency and low temperature the primary driver. (L-68; 1st full para.; right column)

Pg 38 (L-69; 3rd full para.; right column)

4th para - Aug O-ring dia obtained from 18,000 measurements was 0.281 dia with a standard deviation of 0.0013" - no comment just data for future use.

[M21] Pg 39 (L-71; 4th, 5th, and 6th full pares.; left column)

(3) Findings - I'm glad to see these conclusions because I told NASA back in Feb that this was so from my experience.

Pg 55 (L-73; last para.; right column)

Launch temp analysis - info only - all joints predicted local temps of 28 ±5°F.

Pg 56 (L-79; 2nd full para.; left column)

1st para - Pressure actuation testing ...........is maintained - This testing is not applicable to demonstrate sealing down to -10°F.

2nd para - Doesn't this data confirm that low temp, is the major affect on the joint leakage problem. (L-79; 3rd full para.; left column)

3rd para - I'm not sure that these statements are whole truths or even partially true. The data obtained thus far needs to be studied and used for re-design purposes. (L-79; 4th full para.; left column)

Last para - I'm not sure that footprint test data supports this statement. I need to check my data. (L-79; last para.; left column)

Pg 57 (L-79; last para.; left column)

1st sentence continued from Pg 56 - This is contrary to good seal practice.

(C) Findings - I don't know why this statement is significant. Erosion and blow-by seem to occur at random in different joints at different circumerntial locations. I still don't know why temperature alone is not as viable a conclusion or finding as the one stated. (L-79; let full para.; right column)

[M22] Pg 63 (L-81)

This chart says it all - temperatures below 50°F can cause joint leakage. The point that everyone forgets or chooses to ignore is that perhaps 75% or greater of the circumference in any joint contains near nominal gaps (.005 ± .004"). We have always concentrated on what is the minimum O-ring squeeze in any joint as the criteria for acceptance (.020" min) but the actual joints are mostly at much higher squeeze values.

Pg 65 (L-82; let full para.; left column)

Last para - "Prior to the Challenger ..........of the joint" - This statement is not true. MTI & NASA both recognized that putty could hold pressure. In fact tests were run to characterize how much the putty could hold so that the leak check of the seal would truly be a leak check without the putty doing the sealing. That is what established the 200 psi open source leak check to insure that the putty did not mask a leaking seal.

Pg 66 (L-82; 3rd full para.; left column)

Findings and Pg 67 Figure 29 - How can these be interpreted any other way than low temp has a marked affect on the sealing capability of the joint.

Pg 76 (L-85; 5th full para.; right column)

Last paragraph:

"The investigation has shown that the joint sealing performance is sensitive to the following factors, either independently or in combination:

• [M23] Damage to the joints/seals or generation of contaminants as joints are mated
• Tang/clevis gap opening due to motor pressure and other loads
• Mated O-ring compression (static)
• Joint temperature as it affects O-ring response under dynamic conditions (resiliency) and as it relates to forming ice from water intrusion in the joint
• O-ring pressure actuation timing as it relates to putty performance in the joint

I think it is very obvious that an attempt is being made to play down the affects of temperature on the joint failure by these statements.

Pg 77 (L-85; last para.; right column)

Assembly damage/contamination - This whole first para is not supportive of what I have been told that they do at KSC when they mate. The control for mating is not to have a flat on flat condition - the magnitude of out-of-round is a guide period. If all this data was so important a contributor then why doesn't it correlate with all out joint erosion and blow-by data.

Pg 78 (L-86; 2nd full para.; left column)

1st para - Notice how the metal slivers were generated from a slight flat on flat condition which violates the KSC stacking criteria. The 4.1% O-ring stretch is also bull plus the segment sections were not free to move radially as they would be in an actual joint. All this data is highly suspect.

[M24] 2nd para - If this was such a significant problem as it is being mad out to be then why haven't we noticed metal slivers in the zones of eroded O-rings before. To my knowledge, none have been seen or reported with the exception of the DM-5 field joint which failed leak check.

3rd para - If the loads are so low then where is our mechanism for creating slivers at a high frequency of occurrence.

Pg 79 (L-86; 2nd, 3rd, 4th full paras.; right column)

• Gap Opening Statement basically says that max gap opening is 0.029 + .008 = 0.037" at primary O-ring in fwd joint. I measured 0.002 max on the land between the two O-ring grooves. I don't understand how the values are decreasing from what was measured.
• O-ring squeeze as mated (static) - The first paragraph just shows a lack of understanding on how the joints are mated. The final mated gaps on any joint can have no relationship to the ovality at the start of mating. The joints mate thru the mechanism of fixed circumferential lengths and 180 master tooled pin holes.

Pg 79 (L-86; 5th full para.; right column)

Last paragraph - Pure speculation without evidence specific enough to make a good conclusion.

Pg 80 (L-88; 4th full para.; left column)

• 2nd para - This is a very important: set of statements that seems to have a two edge sword. It seems contrary to seal criteria in general. More testing needs to be done to verify or refute this statement.
• 3rd para - Same comments as above
• 4th para - "Joint temperature" - 28 ± 5°F. This is a range of 23°F to 33°F - data only.

[M25] Pg 81 (L-88; 4th full para.; left column)

• 1st para - "The investigation has ...........for ice in the joints" The investigation also showed that it takes more pressure to slide an O-ring across the groove as the temperature is decreased and also the O-ring does not extrude into the gap as easily as the temperature is decreased. Both data references from MTI testing in the extrusion rig.
• 2nd para - I believe our (MTI) resiliency data showing times and joint movement values at first light would be more meaningful. Our data shows that both the primary and secondary O-rings have trouble following the gap opening both with pressure applied and without pressure applied. (L-88; 5th full para.; left column)
• 4th para - Putty performance - Putty sealing potential was recognized prior to 51L. This was the basis for putty testing to determine what pressure to use for leak testing to make sure that putty did not mask a bad seal assembly.

Pg 82 (L-88; 2nd & 3rd full para.; right column)

No one ever addresses the fact that if blow-by occurs (from whatever cause) that the hot gases will severely erode the seals and if they both or only one seals with the remaining seal material that the major cause of the event is low temperature.

• Appendix "A" (L-91)

  Pg A6 - Both pieces of the burned joint have been recovered.

• Appendix "B" (L-109)

  [M26] Pg B4 - The difference in shell deflection and field joint gap deflection between the static and dynamic analyses was negligible óIt was concluded that a quasi-static analysis was adequate for determining gap opening as a function of time caused by various induced loads. info. (L-109; sub para. 4.1; right column)

• Analysis has shown that the case response is dominated by the internal pressure in the motor info only (L-109; 5th full para.; right column)

Pg B-8 (L-111; 4th full para.; left column)

• (C) Response of fully compressed O-ring - Seals do not work this .way. Even our hot fire test with an initial 0.012" gap eventually sealed into the extrusion gap. This section is pure analytical modeling results.

Pg B-9 (L-111; 1st full para.; right column)

• (D) Same comment as above
• Last para - I'm not sure seal touches the groove side walls - check this.

Pg B-10 (L-112; 1st full para.; left column)

• Findings (2) Bull - Nothing to support this statement.

Pg B-12 (L-112; 2nd full para.; right column)

• 1st para - That is, if the primary ...........Least 2 or 3 mils". I thought the referee testing showed less gap opening at the alignment slots.

Pg B-16 (L-113; 5th full para.; left column)

• 2nd para - Many pins are driven into the holes so the final position of the joint is not known but I do believe it stays relatively the same after the weight is removed from the crane.

[M27] Pg B-19

J - Don't agree with this statement at all. (L-114; right column; para. 5.J.)

K- Speculation - if not then why do we see blow-by at a range of temperature (L-114; right column; para. S.K.)

L- More compression should give more stored energy not less - low Temperature simply negates any response of the material. (L-114; right column; para. 5.K.)

Pg B-98

Summary of ambient temperatures: (L-115; right column)

Pg C-44 (L-206; left column)

SRM O-ring stacking damage test. Test description: A section of a flight aft center segment tang and its associated aft segment clevis are secured in rigid fixtures in a tensile test machine. Note: The test is somewhat not valid because the fixtures were "rigid"

Pg C-56 (L-211; 5th item; right column)

Last item explains the hot fire test with 0.015" negative O-ring squeeze (gap) that sealed at 30°F

Pg C-57 (L-211; 6th item; right column)

Not sure that last statement is necessarily true since this test series is a fixed gap test. No one can say for sure if the O-ring would have sealed if the gap had been opening at the full scale motor rate.

[M28] Pg C-77 (L-226; left column)

"O-ring static blow-by test"

The primary objective of the ring was simply to see if blow-by would occur during the ignition transient, especially in the 5 psig to 50 psig pressure region. It was never intended to be a sealing test as such until after the 51L incident because the fixture was already in existence. The results and conclusions don't even mention blow-by measurements above 25°F.

Pg C-80 (L-226; right column)

Results/Conclusions:

Blow-by the primary twice at 40°F - TWR-300152 Pg 6

Pg C-81 (L-226; right column)

".009" initial gap/.027" growth gap"

Blow-by primary but sealing of secondary at 40°F Ref TWR-300152 Pg 7.

Blow-by both seals at 25°F,

Pg C-81 (L-226; left column; "within the limits....")

Last para - Much more testing is needed to understand the statement being made.

Pg C-85 (L-226; right column; "Test No. MTI" 111")

"O'ring Resiliency Test"

Our curves and tables S/B included same comment applied to our dynamic test data. Our tables S/B included this would add only 3 to 5 more pages, but the data is very significant.

• MTI cold O-ring extrusion test results are not reported-TWR-300154.

[M29] This test shows that the pressure needed to move an O-ring into its seated position is affected by temperature. This is a fixed gap test but shows the difficulty of the seal to extrude into the gap as the temp is lowered.

• The following MTI tests were included in the report.
  • Leak check port plug
  • Referee test
  • Full scale joint/subscale dia hot fire
  • 5" hot fire tests
  • Dynamic vacuum putty tests
  • O-ring viscoelastic properties tests
  • O-ring static blow-by test
  • O-ring dynamic blow-by test
  • O-ring resiliency test
  • Frozen joint test
  • Hot O-ring extrusion test
  • Cold O-ring extrusion test with sharp groove edge
  • Ice in joint test
  • O-ring twist test/demonstration
  • Seal contamination tests
  • O-ring assembly damage test

The bottom line is simply - no one has had dedicated time to perform a detail review of all the data. This is an absolute requirement in order to do a proper summary report. Also, this review is required in order to affect the basis of the redesign to avoid making design errors.