NASA STS Recordation
Oral History Project
Edited Oral History Transcript
Interviewed by Rebecca Wright
Huntsville, Alabama – 30 June 2010
is June 30, 2010. This interview is being conducted with Myron (Mike)
Pessin in Huntsville, Alabama, for the STS Recordation Oral History
Project. The interviewer is Rebecca Wright.
I retired, the NASA Administrator decided that he wanted to transfer
all the Marshall [Space Flight Center, Huntsville, Alabama] Shuttle
projects to USA [United Space Alliance]. USA was going to procure
the tank, the motors, and the engines. After I retired, USA came to
me and said, “We’ve got a three-month assignment to help
write a transition plan.” That lasted four years. But some Marshall
personnel weren’t interested in transitioning, so they fought
it virtually every step of the way, and we ended up with a lot of
The Marshall External Tank (ET) Project people didn’t really
want us in their meetings. Lockheed Martin [Corporation] didn’t
want us involved. They wouldn’t let me play solitaire on my
computer so I started writing a history [Lessons Learned from Space
Shuttle External Tank Development: A Technical History of the External
Tank]. This was strictly on my own, not an assignment from USA. The
only USA involvement was a couple of the secretaries would type stuff
for me, unofficially. When USA finally decided NASA wasn’t serious
about transitioning the Shuttle elements they laid off Ken Jones,
who had been the SRM [solid rocket motor] chief engineer, Dennis Godtsen
who had been SSME [Space Shuttle main engine] deputy chief engineer,
Jim Smith who had been SRB [solid rocket booster] chief engineer,
and me. Since I was over 65, I draw retirement pay from USA—if
you get laid off after 65 you draw some of the retirement pay.
I’d finished about three quarters of the ET History by this
time, so I took what I had done, and gave it to the NASA ET project
manager, who was a friend. He gave it to the next generation project
people and they had me finish it. They put me on contract through
Madison Research [Corporation]—Madison is a local firm—to
finish it as a lessons learned document. Fortunately, Madison had
an editor who corrected my poor grammar and a typist who did the typing
for me. That’s how it got published.
Then after the [Space Shuttle] Columbia accident [STS-107], the press
found out about it and wanted copies. Of course, I wouldn’t
give them a copy, so they requested it under the Freedom of Information
[FOI] provisions and got it put on the FOI website. I got calls from
Aviation Week [magazine]; I got calls from New York Times, and the
Washington Post. The Orlando Sentinel came by the house, ABC [American
Broadcasting Company] called me—they wanted to do a special.
By this time USA had brought me back as a consultant on the accident,
and I didn’t want to get involved in talking to the press because
I knew the [NASA] Administrator didn’t want anybody talking
to the press except the members of the board. So I wouldn’t
talk to the press but they said, “We just have a few questions.”
I said, “I wrote it as a lessons learned document for young
They said, “We’re not engineers.”
I was tempted to say, “Obviously,” but I was too polite.
had all those years with the tank, but actually when you started out
you were put with the engines. Give us a brief history about how you
I got out of college in 1953, I went to work as a propulsion engineer
for an aircraft company in Dallas [Texas] making Navy fighters and
cruise missiles. However, my draft board decided I was more use to
the defense effort as a clerk typist in the Army, so they sent me
to a two-year vacation in New Jersey. At the end of that period I
went back to the aircraft company in Dallas. In December of ’58
this company lost two contracts which they had been counting on for
the future so by March of 1960 the company had gone from 20,000 to
8,000. I was 8,001 so I was looking for a home.
When ABMA—the Army Ballistic Missile Agency—research and
development organization split off from the Army and became Marshall
[Space Flight Center], a lot of the people stayed with the Army so
both groups had a lot of holes in their organizations. Chrysler Space
Division was given the task by NASA of finding people for those holes.
So I hired in with Chrysler Space Division, and came to Huntsville
on July 1st of ’60 when Marshall was established. I wasn’t
civil service; I was Chrysler.
Chrysler hired me as a propulsion designer. All my background had
been in analysis and test, and I’m a lousy designer so after
seven months I went to work for NASA. At that time Marshall had responsibility
for all the light and medium launch vehicles, which were Scout, Agena,
and Centaur. I ended up on the Centaur project. Somehow I ended up
as a propulsion engineer, working in mission analysis, trajectory
analysis and performance planning. I spent two years there. We flew
the first Centaur, it blew up. [NASA] Headquarters [Washington, DC]
transferred the program to Lewis Research Center in Cleveland [Ohio,
currently Glenn Research Center]. Lewis offered me a job, but Cleveland
in the wintertime was too cold for a good Southern boy and Lewis didn’t
really want us anyway; it’s just that Headquarters was pushing
So I went down to New Orleans [Louisiana]. At that time they were
staffing up the resident office in New Orleans—and I’m
a native of New Orleans—for the Saturn program. The Michoud
plant [currently Michoud Assembly Facility] was down there. It was
left over from World War II, a 46-acre plant. Chrysler was given half
of it and [The] Boeing [Company] was given half of it, and I moved
into the office that oversaw Boeing.
Just from a brief historical point, the way NASA found that plant
was interesting. We had a guy in our facilities office named [C. L.]
Horton Webb whose job was to search out big facilities that might
be useful for the Apollo program. He went down to this plant. During
World War II, the plant had been active and then had been abandoned.
Chrysler moved in during the Korean War and was making main battle
tank engines. At the end of the war it was abandoned again.
There was one civil servant from Birmingham Ordnance District [Alabama]
and a dozen contractors in this 46-acre plant. So Horton looked at
the plant—40-foot-high ceilings, 125-foot column spacing, had
its own barge dock, had its own runway. He said, “This is great.
Can I have copies of the drawings?” They took him in a room
and rolled up against the wall was every drawing that had ever been
made of the plant. He just sat there and started unrolling drawings
until he found what he wanted, went downtown to a blueprint shop,
got a copy, brought the drawing back to Huntsville.
A couple weeks later, he was down there visiting. The phone rang,
and it was [Wernher] von Braun’s office. He wanted to have a
meeting with a couple of senators and a couple of congressmen. Two
days. So Horton asked the Birmingham Ordnance District [BOD] guys,
he said, “Where’s your conference room?” It hadn’t
been touched in eight years. Inches deep in dust, the halls were this
deep in dust [demonstrates].
He asked, “Well, what about air conditioning?” The BOD
people had a window unit in their office. They had no idea how to
fire up the main air conditioning plant. He said, “What about
food service?” They had an icebox with drinks in it. So he went
out and he hired a janitorial service to clean up the place; he hired
those people with portable air conditioning units that they use for
weddings to stick an “elephant trunk” in the window; he
hired a catering service for coffee and doughnuts; and he hired a
limousine service to pick up the people at the airport.
He did this on his personal credit card. When he got back to Huntsville
and turned in his travel voucher the Finance Office was disturbed.
He said, “Call von Braun’s office.” Von Braun okayed
it. Horton did the right thing. You don’t tell a senator you
can’t meet with him because the place is dirty. To let a contract
to get the building cleaned would have been a three-month job. The
Birmingham Ordnance District guys, they didn’t work for NASA,
so he did exactly the right thing.
At that time Marshall was under von Braun and a very new agency, so
most of the regulations hadn’t been written yet. The philosophy
was to get the job done, that was von Braun’s approach. “We’ll
worry about some of the regulations, some of the purity afterwards.”
I spent eight years there. I was badged to the local office, but I
filled the role of what Marshall called propulsion and vehicle engineering.
Propulsion and vehicle engineering is essentially the mechanical engineering
lab in one of the MSFC technical laboratories. I filled the role of
the mechanical engineering representative. I had been there for eight
years, and by the end of the 1960s the Saturns had finished and all
we were doing was storing them and preparing them for flight. Michoud
had gone from a civil service workforce of 280 down to 35, and it
was still shrinking so I figured I better find a bigger home.
I’d been working with the Shuttle Task Team as a representative
from Michoud and there was a group which was doing systems integration.
They had an opening for somebody in the mission analysis/trajectory
planning world, which is what I had done on Centaur, and I had worked
with those guys directly. They offered me a transfer, and I said yes.
I put in the paperwork, got notified my transfer was approved, but
I was going to work in the engine office. I said, “I didn’t
apply to the engine office.” They said, “You’re
going to work in the engine office.” The engine office had a
representative who was the manager of the Space Shuttle engine in
the task team. I went to work on his staff working engine vehicle
At that time for the Space Shuttle main engine we had three contractors.
These were Aerojet, Rocketdyne, and Pratt & Whitney. The contractors
were finishing up their Phase B study reports. The RFP [request for
proposal] was going out for the Phase C-D. Each contractor had some
differences in their design. On the vehicle Phase B studies, the reports
were open, and all the contractors got to sit in the other contractors’
presentations. The engines were all secret from each other; the engine
guys wanted everything proprietary.
When they established the Source Evaluation Board [SEB] for the Phase
C-D, they took the three Phase B study managers, who were three of
the most experienced people we had, moved them over to the Source
Evaluation Board, and gave me all three study contracts to finish.
At that time since everybody who knew one end of an engine from the
other had gone over to the SEB. I was one of three people at Marshall
who could talk to the contractors. I had all three contractors, and
I had to keep everything secret from each other, and also make sure
that I didn’t tell them anything about what was going on in
the SEB. I was so scrupulous there that if I’d see somebody
from the SEB, I’d cross the street to avoid him. I made sure
I knew absolutely nothing about what was going on.
long did that time period last where you had to be so guarded?
three or four months. It was interesting because one of the contractors
had electric valves, one of them had pneumatic valves, one of them
had hydraulic valves. Pratt & Whitney had actually built a 250,000-pound-thrust
high chamber pressure engine under an Air Force contract that they
had fired a number of times. They had a lot of experience on what
they call high PC [high chamber pressure]. Aerojet had some high PC
experience on the M-1 [rocket] engine that they had worked on years
before. Rocketdyne had no high PC experience so they made a little
1,000-pound-thrust chamber. And this 1,000-pound-thrust chamber was
about this big [demonstrates], and operated at 3,000 psi [pounds per
square inch], which is what SSME operates. All through their proposal
they talked about their high PC experience, which I thought was stretching
things a little bit. Rocketdyne won, but Pratt & Whitney protested
and the program was delayed a year.
Then the orbiter and systems integration contract was awarded. Rockwell
won that but the decision was made that the tank would be delayed
a year to give the systems integration activities time to put the
requirements together. During the time I was working engine [SSME],
somebody got a very smart idea for the External Tank. They said, if
we wait until we select a contractor and come up with his manufacturing
approach and his facility requirements and write C of F [construction
of facilities] budgets, it’ll be three years before we can start
doing any work. So let’s come up with a Marshall in-house manufacturing
plan, Marshall in-house facility requirements, and go in and start
working C of F budgets before we pick a contractor, with the understanding
with the congressional committees that when we pick a contractor if
he wants to build it differently we will transfer the funds or reprogram
them to the new approach. The congressional committees accepted this
as being a smart way to go.
At that time [NASA] Johnson [Space Center, Houston, Texas] had given
Marshall 12 tasks on Shuttle. They established task 13, and since
I didn’t duck fast enough, I ended up heading task 13, which
was to develop this manufacturing plan. Not being totally stupid,
I went over to Marshall’s Manufacturing Engineering Lab and
grabbed about four or five top-notch manufacturing engineers, went
to the facilities office, got half a dozen facilities guys. We put
together a manufacturing plan and C of F budgets for 1973 and ’74.
So when Martin [Marietta Corporation] came on board in September of
’73, we had 14 million bucks [dollars] in the bank for C of
F for them to get started.
Martin was pretty smart in the facilities world themselves in that
they basically hired the Boeing and Chrysler facilities engineering
organizations. Their position was these people have been operating
this plant for 12 years, why bring in a bunch of new people who have
never been here, who don’t know the New Orleans environment?
Obviously with New Orleans, like the Houston environment, it’s
quite different than it was in Denver [Colorado]. They hired most
of these people so their facility requirements pretty much hit the
When Jim [James B.] Odom was given the task of project manager for
External Tank, Jim grabbed me to work the facility requirements and
manufacturing requirements. I moved right into Jim’s office
and got out of the engine world, which was really good for me because
within Marshall there were a lot of very bright young propulsion engineers
who had extensive engine background. My background was more in vehicle
propulsion systems, and really these guys were far better qualified
than I was and this was better use of the people, so I moved into
the vehicle side.
I started on working facility requirements, facility implementation,
manufacturing methods, manufacturing systems. This is in the project
engineering side. Then as various people moved on, I gradually inherited
loads, inherited structures, inherited propulsion, inherited many
of the subsystems, and inherited tooling for a while. Then later on
when we recognized that going to—at that time we had backed
off from producing sixty ETs a year to twenty-four a year. Jim Odom
set up a production office that Jerry [W.] Smelser headed up to go
to twenty-four a year and they took facilities back from me. They
took tooling and they took over the production side, and I stayed
in the engineering side as staff to [Gene] Porter Bridwell.
Marshall at that time had project engineering functions and a chief
engineer. The chief engineer was in the Science and Engineering Directorate.
The project engineering reported to the project manager. I was in
project engineering and stayed in there for a number of years. Then
eventually [Thomas] Jack Lee, who was Center Director at the time,
decided to combine project engineering and the chief engineer’s
office. Jack Nichols, who was the chief engineer at the time, went
over to the advanced solid rocket motor project as chief engineer.
And they combined both offices; they moved me into the chief engineer’s
role but I still had the project engineering function. My staff in
the project office still reported to me but now in the Science and
Jack took most of his staff with him so I ended up with mostly my
project engineering staff now in the chief engineer’s role,
but I had both functions. In project engineering you were more concerned
with the programmatic aspects. As chief engineer you were more concerned
with interfacing with the Science and Engineering Directorate. In
project engineering I usually was the primary interface to JSC systems
integration, and some of the JSC systems integration people I had
the good fortune to work with were outstanding—Dick [Richard
H.] Kohrs particularly. Dick was great. Some of them I was not so
fortunate. I did the external interfaces heavily as opposed to the
chief engineer who was more involved with the internal interfaces
within Marshall. I dealt with NASA Headquarters quite a bit.
When Jim got me in his office, I helped prepare the ET RFP. For the
year gap between the orbiter award and ET, we did something that I
think was sort of unique. We went to the contractors and said, “On
Saturn these were the high-cost items of building Saturns. We now
have to build a program laid out by Headquarters of sixty external
tanks.” We were going to fly sixty a year. We were going to
fly twenty from each of the launch pads at the Cape [Canaveral, Florida]
and twenty from Vandenberg [Air Force Base, California]. The orbiter
turnaround time would be a little over two weeks per orbiter from
the time it landed until the time it flew. That was the program.
There were five contractors who had expressed interest in bidding:
Boeing, Chrysler, McDonnell Douglas, Martin Marietta and General Dynamics.
We went to them and said, “Okay, tell us how you would build
cheap tanks. The only proviso is that nothing will be held proprietary.
If you give us something good, we may write it into the RFP.”
Well, the contractors were pretty open on this, I think primarily
because they felt if we put it in the RFP that would give them a leg
up. So they came in on a regular basis and told us how they were going
to build cheap tanks and all the features. Of course, this helped
me when I was writing the producibility portion of the RFP, because
I had the education they’d given me on producibility.
Martin was quite strong in the standpoint that they had built 300
Titan rockets. General Dynamics put together an outstanding manufacturing
plan, but later they decided to bid as a subcontractor to Boeing so
they were not one of the four bidders we had. My role primarily in
preparing the RFP and the SEB was in the producibility world. When
the award was made I stayed in the production side, and I gradually
moved over to the engineering side as other people moved on.
Regarding “design to cost containment,” our role really
was not so much design-to-cost. The program Level I requirement was
that we build tanks for $2.3 billion per tank. When we presented this
to our Center Director, Dr. [Eberhard F. M.] Rees, he was furious.
He said, “You’re robbing the taxpayers; you ought to be
able to build them for $1 million apiece.” I think he was a
little bit naive on what production costs were.
Von Braun always had a philosophy that he always wanted an in-house
project, so that he had his people knowledgeable about the state-of-the-art
in manufacturing. This was extremely valuable because his people,
when dealing with the contractors, were dealing from a position of
strength. Too often the government, and particularly the Air Force,
have a project manager who’s experienced and half-a-dozen ROTC
[Reserve Officers' Training Corps] second lieutenants dealing from
a position of—I won’t say weakness—but a position
of not a lot of experience. And the contractors can snow them.
At Marshall, we typically had guys who were as strong or stronger
than the contractor. I know the contractors were quite shocked when
we would get started. We would go into a meeting and they would start,
let’s say, in stress analysis. Marshall would show up with a
dozen PhDs in stress analysis. I don’t think Johnson [Space
Center] had that kind of depth. Johnson had some very, very good people,
but they were very limited in numbers. Marshall could really bring
in depth. I’ll explain one area later on where that got very
were a number of areas where Marshall had that in-house experience?
the area of manufacturing, one of the areas that we ran into is you
have a very large tank, and you have to put foam on it. There are
a number of approaches you could use. Boeing on the Saturn rocket
stage—the S-IC—did not have foam, it was just painted.
The S-II stage started off with a filled honeycomb that had to be
bonded in place. It was very time-consuming, labor-intensive. They
went to a sprayed approach, but because of the aluminum material they
used they had to leave the weld lands bare when they proof-tested
the tank. They had to proof-test it cryogenically, so they tested
the tanks at Stennis [Space Center, Mississippi]. They filled them
with liquid hydrogen. Then they had to go in and inspect the weld
lands, after which they had to manually spray the closeout on the
weld lands. I was told by the Rockwell manager from Mississippi that
he had to rent every bit of scaffolding in southern Mississippi.
One of the things Rockwell had started looking at was a “barber’s
pole” approach. You rotate the tank past a spray gun that moves
up, and you just wrap the foam around. Marshall Manufacturing Engineering
Lab jumped on that. There’s an old southern phrase, “We
jumped on it like a chicken on a June bug.” They had a ten-foot
tank, and they’d done a very large amount of development work
on that. We brought the potential contractors in and demonstrated
it to them. We said, “Okay, this is what we’ve done.”
Not telling them that this is what you had to do, but this is what
we’ve done and it makes a lot of sense to us.
We demonstrated that—forming these parts when you have curvature
in both directions, is called compound curvature. Forming these compound
curvature parts, dome gores and ogive gores, is rather complex. Boeing,
on the S-IC, had hydraulically bulge-formed them, that is by taking
a flat plate and with hydraulic pressure force it into a female die.
We looked at that and we looked at explosive forming, which Langley
[Research Center, Hampton, Virginia] had done a lot of. You take the
flat plate and using an explosive charge force the metal, deform the
metal, into a female die. We looked at peen forming where you take
the flat plate and hit it with steel shot and force it into the die.
We looked at stretch forming where you stretch it over a male die.
Boeing had the biggest stretch press in the world. When Boeing did
not win the contract, we asked if they’d take subcontracts and
they were not interested. The stretch press was integral to the 747
[aircraft] manufacturing, and we were going to be pushing it right
up to its limits. Boeing said if we broke the press they couldn’t
make 747s. The 747 was a much bigger profit center than any work they
could get from us. I completely understood where they were coming
from, so Boeing didn’t bid to make these parts.
Martin won the award and went out for competition. A contractor we
had never heard of called Aircraft Hydroforming in LA [Los Angeles,
California] won the competition. They used a stretch press made by
a company called Sheridan Gray, but they modified it to fit our parts.
Our parts were bigger than had been done, and they did a fantastic
job. They became Martin’s best supplier.
This approach of getting the contractors in, having them tell us how
they would build cheap parts, this gave us a good reference to start
the program. Some of the contractors came in and they were going to
spin a one-piece dome. Well, 27 and-a-half-foot means you have to
have a very large plate. You couldn’t buy plate stock that big
so they were going to have to spin it across a weld. But in fusion
welding, the strength of the weld is half the strength of the parent
metal, and spinning deforms the metal. When you deform the metal and
you have a seam down the middle that’s only half the strength.
It’s going to yield more than the rest of the metal, and you’re
going to have some interesting phenomena. They had people who came
in who had not done their metallurgical homework and were telling
us how they were going to do things, and some of them were rather
pleased to leave after we got through asking hard questions. This
was, to me, a valuable tool to bring the contractors in in an open
We didn’t have the other contractors listening, but we had agreed
that nothing would be proprietary. We had Marshall manufacturing experts,
Marshall materials experts, Marshall structural analysis experts sitting
in these meetings, and they were free meetings for the people to ask
questions. It wasn’t the kind of meeting where you sit back
and you just listen. We encouraged our people to ask questions, and
after the first couple of meetings the contractors knew what to expect.
They would bring in their staffs with the right backup. It helped
us to develop a manufacturing understanding.
The four bidders who actually bid were Boeing, Martin Marietta, Chrysler
and McDonnell Douglas. Martin Marietta had to me a big advantage.
They had built 300 Titans so they had extensive experience on building
pretty good size hardware. The Titan was 10-foot in diameter, 80 feet
long. Boeing had built 15 S-ICs. Chrysler had built 15 S-IBs, but
for S-IB the tanks were made by LTV [Ling-Temco-Vought] in Dallas,
and Chrysler assembled the stages. McDonnell Douglas had built a number
of S-IVBs but they again were smaller, and they used a very labor-intensive
technique for insulating.
When the contractors bid, Martin bid this barber’s pole approach.
Chrysler said, “We will put the tank on a horizontal spit like
a barbecue grill and use 50 spray guns and rotate the tank past them.
It is difficult enough to get one spray gun to start. Trying to start
50 simultaneously we felt was unrealistic.
McDonnell Douglas said, “We will build an inside-out S-IVB.”
On the S-IVB, the insulation was preformed tiles that were bonded
in place on the inside with a neoprene liner so that if the tiles
were to break loose they wouldn’t get into the propellant outlet.
They said, “We will make preformed tiles, bond them on the outside.”
When you bond something you have to have a clamping force to hold
it until the adhesive is set. We use a vacuum bag. We put a plastic
sheet over it, seal the edges, and then connect a vacuum pump to suck
the air out, and we have one-atmosphere pressure providing your clamping
force. We call it vacuum bagging. But with hundreds of preformed foam
tiles, they were going to have to do vacuum bagging, and to try and
do that for sixty flights a year, we felt that was unrealistic.
Boeing was going to spray the foam but they were going to spray it
under a shoe, and the shoe was going to smooth it. They’d have
a nice smooth coating, but the foam wasn’t going to be allowed
to free-rise so you were going to have more dense foam. One of the
big advantages of the foam is it’s only a two-pound density
material, two pounds per cubic foot. So with the Boeing approach,
we couldn’t figure out how they were going to maintain free-rise
The Martin approach—they saw what we had done and they fed it
back to us. It’s good business, you can say. You see what the
customer wants and give it to them, but we felt that was a relatively
When the RFP was released, the orbiter tiles had not been finalized.
The initial ET design only had foam on the side walls and forward
dome of the hydrogen tank, and an ablator that Martin had developed
for the Mars Viking lander on the aft dome. This was the heat shield
for the Mars Viking lander. Martin had developed this material called
SLA-561, which stood for superlight ablator.
That’s all the insulation which was applied up to that point
of the program. We had some ablator patches up on the intertank and
the front of the hydrogen tank where the shock impingement came off
of the SRB nose and the nose of the orbiter hit the tank. That was
all we had. We had to eliminate ice. When the tiles were designed,
they reported that an ice cube dropped four inches would crack a tile.
We did not necessarily eliminate foam from coming off but needed to
eliminate ice. That was the big driver. So we covered the LOX [liquid
oxygen] tank with foam and we covered the intertank with foam. Then
the program said, “On all of your protuberances, all the pieces
that stick out that have thermal shorts where they could get cold
enough to freeze moisture out of the air, we want those insulated
also. But you’ve got 500 pounds of weight that you can use.
Start at the front end. Work your way as far back as you can; go for
500 pounds.” That was a rather strong challenge.
Some of the hardware was fairly straightforward, like the hardware
at the back end where the orbiter attaches, the big thrust struts,
the vertical struts. The forward ET/SRB attach fitting, those were
fairly straightforward. The ones that were not—you have to remember
that when the vehicle is sitting on the pad before you start to load,
everything is at room temperature. The orbiter stays basically at
room temperature. The tank initially shrinks to -423 degrees [Fahrenheit].
It shrinks several inches. It’s shrinking with relationship
to the orbiter so the forward attach, which is a bipod spindle, has
to move. Then as you fly and you start putting warm pressurization
gas in the front end of the hydrogen tank, it begins to grow. The
bipod has to move back in the other direction. That says you just
can’t cover everything with foam. You’ve got to have the
capability built in for this to swivel.
Foaming the side walls was straightforward; well, semi-straightforward.
We started off with BX-250, the same foam S-II had used, but it has
fairly poor high-temperature characteristics so we were going to have
to underlay it with ablator. You put the ablator underneath because
if you put it on top, the foam would sublime out from underneath the
ablator and the ablator would then fly off. The ablator is a 17-pound
density and it would definitely do damage to the orbiter.
We had to underlay a lot of the foam with ablator. Marshall’s
Materials Laboratory found a new supplier called CPR. I believe this
stood for Chemical Products Research. They had a product, CPR-421,with
much better high-temperature properties. We started switching over
to it, but then a Professor [Irving N.] Einhorn from the University
of Utah [Salt Lake City] pyrolized a whole roomful. When you pyrolize,
you burn it in the absence of oxygen. He caught the effluent gas in
a cryogenic trap and injected a rat, and it killed the rat. He published
this report. CPR came to us and said, “Okay, we are shutting
We said, “What do you mean you’re shutting down?”
They said, “We’re selling this stuff to insulate steam
pipes in office buildings. If there’s ever a fire and anybody
within a half mile could conceivably have breathed that smoke, we’re
going to get sued once this report is in the literature. So we are
We went in, did some chemical analysis, and decided that the toxic
product is a material called TMPP [trimethylolpropane phosphate].
We said, “Can you reformulate it without the phosphorus?”
They did, and that became CPR-488. We had Southern Research [Institute]
down in Birmingham [Alabama] do an extensive test program of trying
to kill rats with it, and it was proven safe and it met the environmental
restrictions, so we switched over to it.
Because of the environmental requirements for spraying BX250, you
had to be below about 60 percent relative humidity in a moderate temperature
band. CPR was totally different. You had to be below 20 percent relative
humidity. The substrate had to be at about 145 degrees, the foam had
to be at about 150 degrees. The air environment in the cell where
you spray had to be—they actually wanted it hotter than 105,
but because people had to go in there, safety officials wouldn’t
let us go above 105, so we had to greatly modify the facility capability
of spraying it. We actually had to pass hot gas through the tank to
heat the tank wall. Turned out that when we were doing the first tank,
which was the main propulsion test tank, the facility wasn’t
available. So we sprayed it with the BX, which was much easier to
spray, and that was the MPTA [main propulsion test article] which
went over to Stennis. It wasn’t going to see the high heating
Then we had to develop the capability to spray this other foam. We
had a number of issues on spraying foam and on the foam manufacture.
For nonmetallics particularly, you have suppliers and the supplier
buys stuff from somebody, and he buys stuff from somebody, and he
buys stuff from somebody. And it gets down to what we used to call
bucket chemistry, two guys in their garage pouring stuff from one
bucket into another. These subsuppliers have no idea where their product
is going to be used. They have no idea what the effects of a change
The prime supplier, he’s selling us a product but he’s
not really knowledgeable of what the product’s needs are. You
run into great difficulty in the world of nonmetallics. I’ll
give you a couple of examples. CPR got bought out by a major chemical
company, and this chemical company delivered a lot of material, and
when Martin sprayed it, it was the wrong color. So we called the chemical
company. They brought in the vice president of engineering, vice president
of research, vice president of sales, whole bunch of senior people.
We had a big meeting at Marshall. They said, “Well, we used
to allow up to five parts per million of iron. Now we have limited
it to two parts per million because most of our isocyanate is used
in picnic coolers, and the lighter color, if it bleeds through the
liner of a picnic cooler doesn’t hurt. But the darker material,
if it bleeds through the liner people complain.”
NASA was 5 percent of their output, and the other 95 percent was other
sources. So, they tightened down. We said, “Well, we don’t
know what the material that we qualified was since their specification
was less than 5 ppm, not a discrete number. We don’t know really
how critical it is so we’re going to have to go back and test
it.” The test series meant going into the wind tunnels at Arnold
Engineering Development Center [Tennessee] and running a $1 million
test series. The wind tunnels are quite expensive to operate.
They were very sorry about that, but they didn’t offer to pay
the $1 million. Really legally, they weren’t liable. At the
end of the meeting I asked them one question, and I got the only answer
I’d believe. I said, “Are you going to make any more changes?”
They said, “You’re a 5 percent customer. We are not going
to let a 5 percent customer drive our market penetration for the other
95 percent of our capacity. We do promise to tell you in advance if
we make a change, but we’re not going to let you drive us.”
That’s the world that you live in in nonmetallics.
The second problem that we ran into—the primer manufacturer,
the guy who actually made the primers for the S-II, had been making
primers for some time. Martin placed an order and the shipping paper
was fine, cans were labeled properly, the very very preliminary test
that’s done at receiving said the material was fine, and it
was released to the floor. The technician sprayed the 1,000-square-foot
dome and said, “This isn’t the same stuff I’ve been
spraying.” I think he got a bonus for reporting it, hopefully
When he looked at that, they found the supplier had put the wrong
solvent reducer in the can. How do you catch that? In Marshall’s
Materials and Processes Lab, the lab lead was a fellow named Bob Lynn,
an excellent chemist. That was my world in the ET Project Office,
so Bob and I got together. We said, “Is there some way we can
sample this material?” Well, in the days I grew up, organic
chemistry or wet chemistry was titration columns and very very slow,
very detailed, very agonizing work.
Right in this timeframe, industry was developing a bunch of computer-controlled
chemical analysis tools. Bob got with Martin and developed a plan
and we bought Martin several million dollars’ worth of equipment,
each of which would sample for one characteristic and would give you
a spectrum. We called that spectrum the fingerprint. The family of
them, we called a signature, and we fed this into a computer. For
each lot of new material that came in, we compared that to the previous
lots, which would tell us if there was anything different.
Martin hired a very very bright young PhD who’d just gotten
out of the University of New Orleans. She had been developing sampling
techniques for the chemical industry. Located between New Orleans
and Baton Rouge [Louisiana] is just one massive chemical industry,
and she developed the sampling techniques for the chemical industry,
so we set her up in a fingerprinting lab. Dr. Laurie Rando is her
name, and she did a fantastic job. She’s still running the lab,
and prior to the ramp-down she had in the lab employees with two PhDs,
half a dozen masters, and nobody with less than a bachelor’s
degree in chemistry. A group of tremendously qualified people.
NASA Headquarters Safety Office was so impressed, they gave the ET
project $1 million to have Martin write a fingerprinting manual, which
they, NASA Safety, gave to every NASA contractor. ATK [Aerospace Systems]
uses it extensively. I think USA at the Cape does some of that work.
For nonmetallic, we virtually have to do that because there are so
many tiers of suppliers, and the guy three and four tiers down just
doesn’t know what his product is being used for. He’s
selling a product to a spec [specification] which calls out certain
One other material we were using, one of the things we got involved
with when we sprayed foam—the intertank is corrugated, and when
you spray foam on it you end up with big blobs of foam on the top
of the corrugations when you are spiral-wrapping it. So Martin came
up with a plan. They said, “If we manually spray BX-250 between
the corrugations and smooth it, and then put an adhesion enhancer”—foam
won’t stick to itself once it’s cured—“we
put an adhesion enhancer on”—and that was a material called
Isochem—“and then we can spiral-wrap a smooth tank.”
Worked very well for half-a-dozen tanks.
Then suddenly we flew one and we had massive chunks of foam coming
off the intertank. Intertank is normally very benign environment.
It’s not cryogenic, it has a heated purge so it’s roughly
40, 50 degrees Fahrenheit, and it doesn’t see high aeroheating.
It’s a very easy environment for materials. Suddenly, we were
losing massive chunks of foam, big chunks. The umbilical-well camera
pictures from the orbiter looked like the craters of the Moon. What
we found is the Isochem had changed one of their subcomponents. When
they put the material on, it wasn’t completely curing, and then
when the outer layer was sprayed over it—when you spray foam
it’s an exothermic reaction. It generates heat as the foam cures.
That heat caused the Isochem to cure, which gave off gas. So you had
gas pockets between the two layers of foam. What we ended up with
was we found that if the pockets were less than five inches across
they would stay on. If they were greater than five inches they’d
come off. We went in and drilled holes through the outer layer into
the pockets. We didn’t know where they were. We drilled holes
on five-inch centers to vent the pockets so that that way we could
keep on flying.
In this case the vendor made a change—this is again before we
started fingerprinting—we didn’t know the change had been
made, and he didn’t know what the effects of the change would
be. In the nonmetallic, that’s the world that you have to fight.
In metals, there’s a little bit better control.
the fingerprinting process, did you find areas as you went along that
it proved its worth?
not sure what the results were. I know that Laurie ended up building
a much better capability than most of her suppliers. She actually
does the work for the suppliers. They will send in stuff to her to
be tested because she can do better and cheaper with the computerized
equipment. She can run much cheaper tests than most of them could
if they don’t have the computerized equipment. This precluded
the suppliers shipping defective products.
It’s the real world in that nonmetallics are very very—I
won’t say unreliable. But the suppliers, because they have so
many tiers of subcontractors, they don’t really know what they’ve
got. Most of the suppliers just don’t have the kind of money
to completely sample all of their incoming products.
One other point. When we went to the CPR-488, which was the later
version of CPR-421, we applied it to the side walls. It worked great.
One of the tests we ran for the aft dome, which sees extremely high
radiant heating from the solid rocket motor exhaust and the very high
acoustics, was what we called a combined environments test. This was
done at Wyle Labs [Laboratories] just up the highway here in Huntsville.
They put foam on basically a flat plate with liquid helium on the
back side. Using liquid hydrogen gets a little dangerous to get the
cryo [cryogenic] backface. Then you subject it to extreme radiant
heating with heat lamps. You blast it with a horn at about 165 dB
[decibels]. You mechanically strain the metal and then you see whether
it comes off. Well, previously when we’d done it on CPR-421,
everything was fine. When we went to CPR-488, the foam ignited and
burned down to the substrate, and that caused significant concerns.
We didn’t have another material so we came up with a temporary
fix that said the critical condition—the dome is 1,000 square
feet, but the outer portion is curved around the dome where it sees
fairly high airflow velocities so it is only in the center that you
have this burning condition where the airflow is stagnant. For the
800 square feet in the center, when we would build the tank, we would
spray ablator on that 800 square feet and then spray foam over it.
If the foam ignited, it would burn down, but the ablator would protect
the tank from the heat.
The ablator is an interesting material. I mentioned it was developed
by Martin for the Mars Viking lander. It uses heptane as a solvent.
Heptane is white gasoline. Spraying white gasoline out of an atomizing
spray gun is a challenging job. OSHA [Occupational Safety and Health
Administration] Standards say you have to have ventilation that stays
below 25 percent of the lower explosive limit. Martin puts a factor
of two on that and gets it down to 12.5 percent of the lower explosive
In this particular application we had to build a new building so we
could spray this. At a pass we had to spray it 4/10 of an inch thick
at 50/1,000. We had guys out there with spray guns spraying 50/1,000;
then going back with another layer, another layer, building up 4/10.
Everything had to be Class I Division 1 explosion proof. The guys
had to be in full breathing gear. We had enormous ventilation system
in there to stay below the lower explosive limit of the material.
We later found a new manufacturer who could make us a slightly denser—three-pound
density versus two-pound density—material on the aft dome. This
company was called NCFI, North Carolina Foam, Incorporated. Their
primary product was sofa cushions for the North Carolina furniture
industry, but they became one of our major suppliers, and they are
really an outstanding supplier. They are small enough that they are
interested in ET business. The big guys, really they were treating
us as a cross they had to bear. For a large manufacturer, the size
of our orders when we went down to four a year, they just weren’t
interested. When we started out at sixty a year, the whole world was
When we started off on the intertank, Avco [Corporation] in Nashville
[Tennessee] was making it. They have two big thrust panels on the
side that are machined flat, and then have to be formed, then six
panels that are skin/stringer panels. Avco was doing well on the skin/stringer
panels but on the big thrust panels, they broke the first six they
tried forming. So Martin pulled the work back to Denver and modified
the forming technique and was making them at Denver.
Later we moved those. For the next buy, Martin moved them to LTV in
Dallas. LTV made them for a while. When Martin went out for another
buy, LTV said, “Your production rate has just gotten down so
low we’re not really interested. If you can find somebody else,
we’ll work with them to develop the process. If you can’t
find anybody else, we’ll continue to make it.” Which I
felt was a fair attitude.
We went to Learjet, and Learjet made them for a while. Then on the
next buy Learjet no-bid, so Martin pulled it back in-house. The point
is that when you start off with sixty a year the whole world is your
buddy. When you go to these low production rates, finding the suppliers,
particularly from within the big aerospace suppliers, it’s not
economical for them to take these small jobs. They have to clear an
area, maintain an area, maintain the tooling, and the skilled workforce.
The other point is if they do it, and you want to buy enough for twenty
flights, they want to build the twenty flight sets and ship them;
then shut down and go off and do something else. If you’re going
to build four a year, they don’t want to deliver one ship set
every three months; they want to deliver the whole set. At Michoud
we were fairly able to support this because Michoud is a big plant
and we had a lot of storage space, so we did that as much as we could.
One other thing—when we started the program the aluminum industry
was saturated. They were selling all the aluminum they could make,
so they had a rule of thumb that said they would only deliver what
you had bought the previous year. Of course when we came in, we hadn’t
bought anything the previous year. Fortunately, Martin had a little
bit of clout and was able to get the aluminum industry to deliver
materials for us. But 2219, the alloy we were using, is 6.4 percent
copper, which from an aluminum refining process poisons their furnaces.
So they would only run 2219 once a year. They would collect all the
2219 buys, deliver them once a year, and then shut down, clean the
furnaces, and go back to making other aluminums.
We did run into one other interesting problem with Reynolds [Metals
Company]. When you make the aluminum plate that we use, which is age-hardened
and work-hardened—after the plate had been rolled, it is solution-heat-treated—that
is you put it in an oven at 985 Fahrenheit to drive the alloying elements
into a solid solution. Then it comes out of this 985F oven and has
to be quenched in ten seconds. It goes through a water spray bath
with massive water sprays top and bottom.
Reynolds ran into a problem where the lower spray nozzles were clogged
so they were getting inadequate quenching. The way this was found—General
Dynamics was building F-16s jets at the time. They were selling them
to Belgium, and as part of the offset Belgium was going to manufacture
some of the pieces. Well, they sent the material to the Belgians,
and the guys manufacturing it, who were pretty knowledgeable—I
don’t know whether they were using manual machinery as opposed
to automated machinery or what, but—they said, “Something’s
wrong with this material, it’s too soft.” So they went
back and checked. Reynolds had delivered probably hundreds of tons
of soft material to the whole industry. There was this massive exercise
looking for soft aluminum by DoD [Department of Defense], commercial
airlines, and NASA.
We were lucky. When we would get our plates, the plate would come
out flat, and the machining subcontractor would machine the upper
surface to get it absolutely smooth. Then they would turn it over
and they would then machine the Ts into the plate. Because there was
no need to rotate the plates, for us the soft part was on the bottom
so when they turned it over, they were machining away the soft metal.
So we lucked out.
A number of the other customers had major major problems. We had a
team out chasing soft aluminum for six months. Marshall had an aluminum
metallurgist, fantastic lady, named Hap Brennecke. Her name was Margaret,
but she went by Hap. She had been a process metallurgist at Alcoa
[Inc.] on 2219. Alcoa developed it for ABMA. ABMA had hired her, and
Hap became Marshall’s senior aluminum metallurgist. Hap had
a unique personality. If you were humbly ignorant, she would spend
days teaching you aluminum metallurgy, more time than you could possibly
tolerate. If you were arrogantly ignorant, she had a tongue like a
rapier and she would carve you to pieces. We had contractors come
in who were arrogantly ignorant, and they were not pleased dealing
with Hap. She spent six months chasing soft aluminum.
When the ET RFP was released, the orbiter tiles had not been finalized.
In addition to the bipod, when you start to load liquid oxygen, the
feed line immediately gets cold. It goes down to -300F. It shrinks
with relation to the hydrogen tank, which is warm. Then as the hydrogen
tank gets cold, it’s 125 degrees colder, so it shrinks with
relation with the feed line. As it warms up, it grows. All these brackets
have to provide movement capability so you can’t just encapsulate
them in foam. The cable tray stays at room temperature but the hydrogen
tank shrinks. All these pieces have to be able to move to provide
some sliding capability or some movement capability.
We came up with designs on those. The designs, if done properly, would
hold. Foam wouldn’t come off. But one thing that we, Marshall
engineering, Martin engineering, didn’t do was a technique called
process deviation analysis. You look at a manufacturing process and
say, “Okay, if we deviate from this process, this step-by-step
operation, if we deviate, what is the implication?” Then, “How
do you detect that?”
NDE [non-destructive evaluation] on foam was virtually impossible
because X-ray goes right through it. The technique North American
used on the SII stage was one quality supervisor that used a half
dollar coin; he’d go around tapping with a half dollar. We tried
to automate that system. We actually had a thumper that would go in,
and we would get the spectrum of the sound response and analyze it,
and we tried using thumpers. Now they have developed this technique
using the X-ray that they use at airports. But even that, that’s
still a raster system. It’s very very slow. So really NDE for
foam is very very difficult.
What we didn’t do—and I say we, Marshall engineering,
we, Martin engineering—we didn’t really understand the
process failure modes. We, Marshall engineering, Martin engineering,
and JSC didn’t understand the failure effects of foam coming
off. Because as you remember after the Columbia accident, the program
manager at Johnson says, “Oh that little piece of foam couldn’t
possibly have damaged the wing.” You may have seen that on television.
All of us didn’t understand the risk, but as far as the process—we
certified spray technicians. We certified them to spray, and we gave
them the design and told them to go spray it, but we didn’t
look at where the process could fail the design. The classic example
is, if you left a void in the foam, that void trapped gas at one atmosphere,
14.7psi. As the vehicle would climb out, the pressure in the void
would act on the projected surface area. That pressure-area term gave
you a certain force. If that force were greater than the shear strength
of the wall around it, it would pop a chunk loose. If it were less
than shear strength, you’d fly and never know it.
We didn’t go into the intensive analysis of the process and
then the training of the technician to ensure they wouldn’t
leave the voids. After Columbia they did of course, but this basically
was the cause of the Columbia failure—and I’m not faulting
the technicians. I’m saying the technicians had not been trained,
had not been sensitized to the risk in their process. It was engineering’s
weakness. Really Marshall engineering, Martin engineering, and to
a degree Johnson, because when we dealt with the Johnson guys foam
coming off was always a refurbish issue for the orbiter. It was never
Two flights before Columbia, we lost a big chunk of foam. The ET project
manager sent his deputy to the Program Requirements Change Board,
Level II—Neil [E.] Otte was his deputy—to open an IFA,
in-flight anomaly. The program said, “We don’t need an
in-flight anomaly, we know about foam coming off.” The program
manager turned to the ET project manager and said, “Do a study
to see what you can do to improve it.”
The project manager went one step beyond that. He directed Martin
to submit an engineering change to fix it. Martin’s chief engineer
had just retired because, when Martin merged with Lockheed, the retirement
system was such that a whole bunch of their people had retired in
December, right before the Columbia accident. The chief engineer had
just retired and he was brought back as a consultant to head up this
study of redesigning the bipod ice protection system. But because
we didn’t realize the risk, and because Johnson apparently didn’t
realize the risk, we treated it as a routine engineering change, not
as a “stop the world we got to get off,” not as a “ground
the fleet” type of change. And that change was in work at the
time of Columbia.
In the area of feed line brackets—cable tray brackets, the bipod
spindle—all those were difficult because they had to move. When
we started, the program we had a choice. We could have gone to an
active control system—that is, heaters—or a passive system.
The active system would have meant that we would have had to power
high current flow coming across from the stand the ground, through
the intertank into these heaters. According to the Cape, we would
have had to add wiring in the tower to get this high current flow,
because at the ET umbilical we had no high-power circuits. We would
have had to add instrumentation. The instrumentation would have been—in
those days had to be hardwired all the way back to the Launch Control
Center. It would have been tied into the launch processing computers,
because laptops and PCs [personal computers] didn’t exist.
Or we could go to a passive system, cover it with foam. We put a calrod
heater in the spindle itself, but that was it. That didn’t have
any instrumentation attached to it, it was just a heater. We chose
the passive system rather than the active because of the problems
with an active system. Now they’ve converted and they’ve
gone to an active system, of course.
The decisions were made, and the reasons were that the passive system
was much easier, not just for ET but for the whole stack, the whole
program. The issue of foam on the protuberances has always been a
challenge. We felt that there were certain areas where you just couldn’t
perfectly seal it, and we would get some small ice issues. The program
basically was accepting those. Since then they’ve come back,
done redesigns, and fixed those. But the whole issue of foam and orbiter
tiles has always been a challenge.
you say the other challenge would be the weight component? That every
time you designed it affected the weight of the tank?
Every pound of weight on the tank is a pound of orbiter payload. I’ll
go into some of the weight reduction programs we went through and
discuss the different manufacturing processes and challenges related
Delay of aluminum. I mentioned that the aluminum industry would only
let you buy the previous year’s supply. One thing Martin has
done that to me was very smart, they would go in with a yearly buy
of the aluminum for all of the aluminum used at all locations, and
then have the aluminum mills drop-ship it to the various suppliers
who are going to be machining it, rather than for each guy go in for
his small buy and the mills having to run off a batch for this guy,
batch for that guy. Martin goes in there, they identify all the different
forms and shapes—whether it’s a plate or sheet, or casting
or forging—they buy it under one buy.
Of course that means Martin procurement people have to keep track
of all of this. They have to know what the suppliers need, when the
suppliers need it, and get it into their order. And it’s a yearly
order to be drop-shipped to the suppliers. It did work out well in
that we had a pretty good response out of the aluminum industry. When
you buy from jobbers (supply houses), traceability can be lost.
Regarding various tests and their uniqueness, we had two classes of
tests. One was the overall system test. The other was a component
test. From a systems test standpoint, there were two classes. There
was the one that was ET-unique and the one where ET was just part
of a system. The ET-unique were the structural tests. We tested the
liquid oxygen tank with barium sulfate. Barium sulfate is drilling
mud, which I’m sure you have heard of. When you have liquid
oxygen at a 70-pounds-per-cubic-foot density at three Gs [gravity],
it gives you the equivalent of 210-pounds-per-cubic-foot material
pressing on the back of the tank. To do this we used barium sulfate,
We tested the LOX tank at Marshall in Building 4619. During the room
temperature test, it tested fine. We tested the intertank at Marshall
in Building 4619. We had big hydraulic jacks. We could load it, putting
loads into all the different interface points. We tested the hydrogen
tanks under the legs of the old S-IC test stand. We actually used
the legs of the test stand as the load frame to react to the loads
we’d be putting into the tank. We took the tank to 140 percent
of design load for three different load cases filled with 53,000 cubic
feet of liquid hydrogen and went to 140 percent or zero design margin,
three times. You sit well back, because it was a challenging test.
The tank passed fine.
We learned one thing, and this was later something that gave us a
problem at the Cape. When we have the tank sitting between the two
solids and we load hydrogen, the tank shrinks radially. So that aft
dome is tending to shrink. Because the SRBs are anchored at the top
and at the bottom they’re stiff, this causes a tension load
to go into this aft dome as it shrinks, causes the dome to go ellipsoidal.
This ellipsoidal dome tends to deform and pop foam off.
We said, “Okay, where do we go from here?” We’ve
got the hardware built for the first four, five vehicles. For the
first six flights down at the Cape when we would stack the SRBs, we
put the tank in place, attach it at the front end, and then using
giant hydraulic jacks pull the SRBs apart, physically bend the SRBs—because
they were fixed at the top and fixed at the bottom. We would bend
the SRBs, put the tank in place, snug up the struts, and then release
the load, which would put the tank in a precompression state so that
when it went into tension from the shrinkage it would have to take
up those compressive loads before it went into tension. We flew the
first six vehicles that way.
None of us were overjoyed about bending SRBs. Fortunately we had stopped
doing this before the [Space Shuttle] Challenger [STS 51-L accident].
If we had been bending SRBs, [Morton] Thiokol [Inc.] would have never
accepted the fact that that was not the cause of the Challenger accident.
But we had stopped it well before the accident—we went in and
we added 800 pounds of aluminum to the ET aft LH2 [liquid hydrogen]
dome to stiffen it up. That was the structural test program.
On the main propulsion test, that was at Stennis. An awful lot was
learned there, but I think if you talk to the propulsion guys they
can probably give you a better handle there. I talked to Len [Leonard]
Worlund recently. Len was chief of propulsion analysis for years and
years and years in the lab, and then later he was chief engineer for
SSME. He took over after Otto [K.] Goetz, an extremely competent individual,
retired. Former college professor and on the MPTA he’s extremely
knowledgeable. One other person you may want to talk to on MPTA here
in town is a guy named Jim Bruce. Jim was one of Rockwell’s
senior managers at Stennis on MPTA. He’s up here now but he
was a Rockwell employee. Very knowledgeable on what was done on MPTA.
I think those guys can talk MPTA better than I can.
The last test we ran was called the Mated Ground Vibration Test. This
is where we stack the whole vehicle and shake it. We stacked it with
one set of loaded SRBs. I say loaded—they were inert grain—and
one set of empty SRBs. The loaded ones were to simulate the launch
condition; the empty ones are to simulate the end of flight condition.
What you’re looking at is bending frequencies, bending nodal
crossing points. What you’re trying to do is verify your analysis.
Your analysis predicts how the tank is going to flex; this test verifies
There were two interesting points to me on that one. Point one is
we used the Enterprise. They flew it into Redstone [Army Airfield,
Alabama], which is a very short runway, on top of a 747. We all got
out there and watched. It was interesting. Before we handled Enterprise
we wanted to make sure we could handle it safely; that when we took
it off the airplane, put it on the road, our transporters, that when
we lifted it into the stand—so we made a simulator which was
just angle irons and steel. It matched the Enterprise overall dimensions
and the weight and the cg [center of gravity], to make sure our handling
techniques were safe.
After it finished its role, we put it in the boneyard. A few years
later the Japanese came in. They were going to have a space fair,
and they wanted Enterprise. Well, the Smithsonian [Air and Space Museum]
had Enterprise up in Washington [DC]. They weren’t about to
turn it loose. [The Japanese] wanted a flight orbiter. Well, your
friends at Johnson weren’t about to turn them loose so they
said, “Well, what about the simulator you had?”
We said, “Doesn’t look anything like an orbiter.”
They said, “Make it look like an orbiter.”
We said, “Well, you can do that in Japan for a tenth of the
They said, “No, we want something that was used in the program.”
So they hired [Teledyne] Brown Engineering [Inc.], a local firm in
Huntsville, to put a fairing over this whole thing, make it look just
like an orbiter, shipped it to Japan for their space fair, shipped
it back. When you look out in back of this building in the U.S. Space
& Rocket Center that is the orbiter on display—that’s
where that orbiter came from. The other thing we learned on MGVT [mated
ground vibration test] —as part of the test, we loaded the LOX
tank with water. LOX is 70 pounds a cubic foot, water is 62.4. It’s
pretty close. It’s close enough to give you a good test. You
can use what they call a Prandtl number to scale it.
When we loaded it, we had the vent valve open. We are just filling
it, pouring water in a bucket. The ogive portion of the LO2 tank buckled,
and we pulled a buckle about six feet long. So we said, “What’s
going on?” We went back and looked at Martin’s model.
The model said it should be stable under this condition, the weight
of the water.
We sent a half dozen PhDs down to Martin, and they dug into the details
of the model, and discovered the model had been hardcoded at 1.7 psi.
The model knew that there was 1.7 psi pressure in the tank, even though
when they ran it they input zero pressure. We went back and corrected
the model, and the model said it is supposed to buckle. This is what
was happening—when we had performed the qualification test of
the LOX tank, we had tested it on the floor on a solid support. But
when it was sitting on top of the intertank, the intertank has a big
crossbeam. There are two hard points on the intertank. The rest of
it is relatively soft. The weight of the water caused the parts that
were relatively soft to push down, causing a tension load on this
ogive, causing the ogive to buckle.
We put a requirement on the Cape that said, “Until we can fix
this, always keep a minimum 1.7 psi in the tank any time there is
any liquid on board.” The Cape was able to do that, but one
of the things you like to do when you’ve got propellants loaded
is open the vent valve and let the propellant boil. As it boils it
carries off heat and gets denser, and then you continue to replenish
the material that is lost. It’s called densification of the
propellant. You do this on both the LOX and the hydrogen side. Well,
we couldn’t densify the LOX because we couldn’t leave
the vent valve open.
For the first six flights we did not densify, but Martin on the lightweight
tank redesign went ahead and put 200 pounds of aluminum in the front
end so that we could take the loads. This let us go back to densifying
the propellant. Even now between 2 percent and 98 percent we have
to keep pressure in the tank, because to stiffen up the whole tank
would have cost 800 pounds. Working with the Johnson Systems Integration
Team, we felt that the problem for the Cape of keeping 1.7 psi in—which
is fairly straightforward for them—between 2 percent and 98
percent, was an acceptable issue to gain 800 pounds. We went in on
the lightweight tank and actually added 1,000 pounds of aluminum at
the ogive and at the back end to fix problems, and still took 10,000
pounds out of the ET, which is 10,000 pounds of payload. But we had
to add 1,000 pounds of aluminum back there because of these things
that we learned.
The component tests, like the big structure that is at the back where
the orbiter attaches to the ET, all that was tested at Michoud. We
set up all sorts of text fixtures at Michoud. I don’t know of
anything we really learned uniquely on that. One thing we did learn—we
had excess margins in many cases. We were more than the 1.4 factor
of safety. When we went to the lightweight tank [LWT] one of the things
we did, we went back and scrubbed this excess margin out because the
program requirement is a 1.4 factor of safety.
One other thing we did on the LWT that was unique, we said, “Factor
of safety is there for four reasons.” One reason is because
the uncertainty in material properties, but we were using MIL-HDBK
[military handbook]-5 “A” values. “A” values
is where you have 99 percent probability, 95 percent confidence that
the material is at least this strong—so we did not have to maintain
a big margin for uncertainty in material properties. Next one is uncertainty
in the ability to analyze the material, the structure. Well, we had
a very very complex structural analysis and test program, so we said,
“We don’t have to maintain this high a factor of safety.”
The third one is variations in the manufacturing capability. All the
tanks were proof-tested, and I’ll go into that later on.
The fourth is uncertainty in induced environments—airloads,
winds, that type of thing. So what we said on the lightweight tank,
we would maintain a 1.4 factor of safety on the less well behaved
loads but use a 1.25 factor of safety where we had well defined loads,
which are thrust loads, pressure loads, inertia loads. Where we had
less well defined loads like aerodynamic loads, wind loads, gusts,
shears, we would maintain the 1.4. That was a break in the industry,
but it let us take some weight out.
Another weight reduction we had—the pressurization system, which
is controlled on the orbiter, was such that a single failure would
get us outside of our control band. Dick Kohrs had the MPS [main propulsion
system] group go in and redesign the pressurization system so that
it would take two failures to get outside the control band. This let
us proof-test the tank to the top of the control band rather than
the top of the relief band, which is 3 psi higher. On a tank this
big, 3 psi is a lot of weight so we were able to take some of that
weight out. There were some new materials that had come along, new
titanium alloys and new aluminum alloys for the secondary structure.
Where we had demonstrated excess margin, we went in and took the margin
There was a fellow named Frank [S.] Boardman, who was Marshall’s
structural analysis chief for ET. Frank and I put in a requirement
that wherever we ran a test we would test to failure rather than just
test to 1.4, or test to failure or the limit of the test equipment
because that told you what your margins were. On the Shuttle program,
the loads seemed to change every other week. I talked to our loads
chief at Marshall who had grown up on the Saturn program. He said
if we had deliberately set out to pick a configuration as hard to
calculate loads as it could be, we could not have done a better job
than we did on Shuttle, particularly with the orbiter wing and tail
being so sensitive. Johnson fine-tuned everything around the orbiter
wing and tail. Whatever else fell out we had to live with, so ET was
getting loads changes on a weekly basis. By testing to failure, we
knew what our margins were, and that helped significantly.
We later were able to come up with load indicators which Johnson loaded
into their prelaunch computer, where we were able to tell them exactly
what the tank capability was. If the upper atmosphere winds were so
high that the load indicators were exceeded, we could not fly that
day. We had several hundred load indicators loaded into the Johnson
day-of-launch calculations. As you know they launch balloons and they
calculate loads based on those balloons. With these load indicators,
we could tell very clearly whether the tank could stand it or not.
Prior to that, we had to have a team of stress analysts and loads
experts standing by at Martin for every launch. They would sit out
there every night, and when the balloon data came in, they would have
to analyze it. That was a waste of money.
Of course the tank was involved with range safety. The Air Force range
safety people are chartered by the President. Their goal is to make
sure nobody on the ground is injured. When John [W.] Young flew Shuttle
[STS]-1, he and [Robert L.] Crippen had ejection seats. John has said
many times that the agreement he had with the range is when we took
the ejection seats out, they were going to take the range safety system
off the tank. The range did not do that. They reneged, or Young claims
they reneged. I think they claimed they never had that agreement.
It was probably fifty or sixty flights later before we took off the
range safety system of the ET. Range safety is propellant dispersion;
it cuts the two tanks open and dumps the propellant. Since the tank
has no propulsion system of its own and it just goes along with the
orbiter, Young was quite adamant that the tank did not need a propellant
dispersion system. The solids do, of course, but the tank didn’t.
Young was adamant about that and he fought that for years and years
and years and finally won.
From a tank standpoint, the only thing we furnished on the range safety
system was the linear shaped charges. The command receiver/decoder,
the batteries, the antenna, the confined detonating fuse—all
that was GFE [government furnished equipment] to us from the SRB project.
We installed it, but as far as saving money on the tank, it was not
part of the tank.
Regarding what was planned to determine the extent of the debris,
I wasn’t involved in that on the first flights, but when we
went to lightweight tank, we had to go back and redo it. It was interesting.
I was Marshall ET representative to the Range Safety Panel. The way
the program works, Program Level II calculates the ascent trajectory
and the ascent heating. They turn that over to ET. ET then calculates
the reentry trajectory and they calculate the break-up altitude. I
think Martin had a good background on this from their Titan ballistic
missile reentry. They then come up with a debris catalog which shows
how the tank breaks up and gives you a random scatter of the size
and shape of the pieces, of the ballistic coefficients, of the lift-to-drag
ratio, of the velocities, and the direction.
On the super light tank we said, “Okay, let’s do a worst-on-worst
analysis. Take the worst-case velocity, worst-case lift-to-drag ratio,
worst-case ballistic coefficient, go in one direction. And then the
least, go in the other direction, and let’s see what the footprint
is.” The intact impact point, the point where it would have
reentered intact, was south of Hawaii. So with the intact impact point
south of Hawaii, the nose of the footprint was south of Lake Charles,
Louisiana, and the heel was south of Borneo. It flew clear across
Mexico, clear across Texas. That wasn’t what we wanted.
We said, “Let’s do a probabilistic assessment,”
and went back and did a Monte Carlo type probabilistic assessment,
and ended up with a footprint about 1,600 miles long and about sixty
or eighty miles wide. Range has to approve the analysis technique,
and then Shuttle has to do an analysis before each flight and present
it to the range, and the range has to approve that before they can
fly. There’s an international treaty that says you can’t
dump anything within 200 miles of a foreign landmass or 50 miles of
a domestic landmass in the event of a no-failure. In the event of
a failure if it hits in the surf, it is okay, but if it hits on the
beach you violated an international treaty.
As far as what actually hits the ground, the extent of debris—during
the second mission, the second Shuttle flight, that’s when they
were coming down in the Indian Ocean. There was a tracking ship in
the Indian Ocean to look for pieces, but it was a Navy ship and they
were charging by the day. The program kept slipping so Level II canceled
the tracking ship. We have never had any hard data as to what actually
hits the water, unless there is a Russian submarine or Russian trawler
out there and they haven’t told us. But we have had data from
a Navy Cast Glance airplane watching the tank reenter; we’ve
had Air Force tracking aircraft.
We have NORAD [North American Aerospace Defense Command] data that
tells us the altitudes at which it breaks up. It breaks up in three
phases. We have two reports from a South African airline pilot flying
from South Africa to Australia who had been vectored off of the flight
path by the notice to airmen and mariners. He saw the sky filled with
glowing green pieces so he circled the airplane so all of his passengers
could see it. This happened on two flights. He wrote us a very nice
letter. But as far as what actually hits the water we have never had—as
far as debris, we don’t know.
There are seven titanium castings, which I’m almost certain
survive. The big aluminum forgings where the SRB attaches at the front
end, they survive in some form. Some of the other big aluminum structure
where we tie the orbiter, I’m almost certain that will survive
in some form. Some of the propulsion parts which are stainless or
21-6-9 steels, they I’m sure will survive in some form, but
as far as having any hard evidence that’s all conjecture.
tank was always envisioned as a throwaway item?
yes. To try and recover the tank would be extremely expensive. It
is a thin-wall aluminum, you have to get it down to extremely low
velocities to keep it from breaking up when it hits the water. In
that case you would then still have an enormous refurbishment cost
after it has been dumped in the ocean.
SRBs are a different animal. Remember, the tank operates at about
35 pounds per square inch internal pressure, so its wall is roughly
100/1,000 of an inch. SRBs operate at 1,000 psi. They are D6AC steel
and the walls are quite thick, so they can take the water impact loads
with parachutes. The tank cannot take the water. You have to get it
down to virtually zero velocity, and that means in addition to parachutes
you probably have to have a retrorocket pack, you’d have to
have a sensor system, you’d have to have some sort of a guidance
system to make sure the rockets were pointing down. It gets expensive.
When I was chief engineer, Ken Jones was the chief engineer for the
motor and Jim Smith was chief engineer for the booster. We used to
have discussions as to, is it worthwhile trying to recover the motors?
Jim’s comment was if we had started from scratch not trying
to recover them, we wouldn’t have had to spend the money to
develop parachutes, wouldn’t have to buy the recovery ships,
wouldn’t have to refurbish the motors and spend all the refurbish
expense. Jim said it’d be a push. His judgment is we would not
really be saving anything by recovering. This was a political ploy.
Ken is different. His philosophy is that by having the hardware to
look at he feels that we may have detected some incipient problems
that could have caused issues later. He likes the idea of looking
at the used hardware. On ET, we would have loved to look at the hardware.
Martin put in a proposal that instead of immediately staging a tank
off to take the tank to orbit—and it’s 50 feet per second
short of orbital velocity at staging which would be a very small payload
hit—and have astronauts go EVA [extra vehicular activity] and
look at the tank insulation.
Eagle Engineering [Eagle Aerospace, Inc.] from Houston did a study
for us. This was Owen [G.] Morris and some of his guys. It was doable
but the Shuttle Program Office was not interested and that raises
an interesting point. President [Ronald W.] Reagan issued an executive
order that said that NASA would take the tanks to orbit and turn them
over to industry if industry would come up with a good plan. We put
an ad [advertisement] in Commerce Business Daily and had ten responses.
The responses were everything from a consortium of 57 universities
to a junior high school science class.
The consortium of 57 universities called UCAR, University Consortium
for Atmospheric Research, is headed by the University of Colorado
[Boulder]. It includes MIT [Massachusetts Institute of Technology,
Cambridge], Caltech [California Institute of Technology, Pasadena],
and all the biggies. Then there was an outfit called Global Outpost
[Inc.]. These were the two that went forward.
The requirements were they had to pay $50,000 up front. They had to
pay all of NASA’s expenses. Both of them let contracts with
Martin to look at taking the tanks to orbit. When the time came for
them to put up cash, both of them backed down. That executive order
is still out there I guess. The junior high school science class response
graded 6th or 7th, and theirs was spelled correctly.
One thing that you’ll find interesting, if you remember, Skylab
started as a wet workshop where Marshall was managing it because we
were going to take an S-IVB to orbit and outfit it in orbit. Then
they came up with a plan that said we would outfit it on the ground.
Well, Johnson took the position that this should be their job because
it is a spacecraft. Because Marshall was using an S-IVB, Marshall
stayed on as manager of Skylab. And Johnson has never forgiven us;
any time we come up with an approach of taking a tank to orbit, Johnson
sees Skylab, or “The Skylab Syndrome.”
There is a very bright young professor from Smithsonian Astronomical
[Observatory, Cambridge, Massachusetts] who wanted to take a tank
to orbit and convert it into a gamma-ray telescope. Since it was being
done in Alabama he was calling it GRITS. Marshall’s advanced
studies guys worked with him on it quite a bit. Again, he didn’t
have the money but that was completely doable.
When the Shuttle became operational—that was the sixth flight—Johnson
recognized they needed more performance so they gave ET an action
to reduce the weight of the ET by 6,000 pounds. orbiter was given
a task to reduce the weight by 4,000. I think on [OV]-103 [Discovery]
they reduced it by 2,800 and maybe on 104 [Atlantis] and 105 [Endeavour]
they got to 4,000. Of course on 102 [Columbia], they didn’t
go back and attempt to rework and 099 [Challenger].
As I mentioned, we developed the lightweight tank. I mentioned what
we did was to take the weight out of the ET. This was a very effective
program. I was happy with it because I negotiated the nonrecurring
cost in the contract at $45 million; it came in at $43 million. We
delivered a 10,000 pounds lighter tank as opposed to a 6,000 pounds
lighter tank requirement, and we delivered it one day ahead of schedule.
Everybody said, “ET is no problem, don’t worry about it,
Some other challenges we encountered—the stiction. We have ullage
pressure transducers which are pot wiper transducers, there’s
a potentiometer with a wiper over it. The early ones would stick.
We went into a slow ramp rate test at acceptance and solved the stiction
problem. Not solved it, just threw out those that were sticky. Stiction
stands for sticky friction.
Transducer output dropout—when the ET was sitting on the pad
and just holding pressure in the tank, the ullage pressure transducers
(the pot wiper ones) would tend to drop out, but as soon as you came
off of that pressure they’d go back to working normally. Only
one flight did we ever have a flight issue. There are four transducers
with only three active during flight. The orbiter computer would check
them very late in the flow, and if one of them was bad, it would switch
in the spare. Only one flight did we ever have to switch in the spare,
and that was interesting.
On the FRF [flight readiness firing] for 104, during the FRF we had
three transducers acting up. Since it was an FRF we could pull them
out and send them back to the lab. Tested them every which way you
can imagine, examined them—they were from different lots; they
weren’t all the same lot. We could find nothing common about
them. We put new transducers in that orbiter, and on that flight one
of the transducers dropped out, and the orbiter computer switched
in the spare. Never had it happen before or since.
but it worked the way it was supposed to.
got little sense hoses, different length sense hoses. We looked. “Could
you have swapped the sense hoses? Their orifices? These dampened any
oscillations in the pressure. Could the orifices have been swapped?”
They looked at everything. Since we had the hardware we could examine
it in depth. We did not find anything wrong, but it was only on that
first flight of orbiter vehicle-104.
to the lightweight tank.” The paint was on there—there
was some concern that ultraviolet radiation would damage the foam,
so we painted the first tanks with a white UV [ultra-violet]-protective
paint. We put a bunch of panels up on the roof of the factory at Michoud,
let them sit there for three or four years, and also had some panels
at the Cape exposed to the Cape environment.
After we’d had a chance to analyze those we found that the damage
was a few angstroms deep. There’s 7,000 angstroms to an inch.
So the damage was insignificant. By taking the paint off we were able
to save about 400, 500 pounds and a lot of labor, because you don’t
paint something that big casually. So we took the paint off. The Public
Affairs Office was unhappy; they preferred the pretty white ET but
it wasn’t worth 500 pound as payload.
“Challenger accident.” From an ET standpoint, we were
relatively unhit. One of the things I did on lightweight tank that
Mr. Boardman insisted I buy for us was a stress analysis report. On
lightweight tank we had to redesign basically every part because we
took the margins down so Martin was going to have to redo their stress
analysis. The initial stress analysis, they had the job shoppers come
in from all over the world, and each one documented the stress analysis
they had done on their part in their own format and had it in notebooks,
loose-leaf binders, every way you can imagine. When we went to the
stress analysis report we put everything in a common format and had
it all in a 17-volume document.
Because Martin had to redo all their stress analysis, I was able to
buy this for $50,000, which was a tremendous saving. After Challenger
Mr. Boardman took the 17 volumes and got two of his stress analysts
in the lab assigned to each volume and sat them down and said, “Is
there anything in there that could conceivably have caused the accident?”
They scrubbed it, they finished that in a week. The programs which
didn’t have stress analysis reports I’m sure took months,
but we were able to get that done immediately.
Also, in manufacturing you write rejection tags. When a rejection
tag is dispositioned on the factory floor, it can be dispositioned
to scrap up to a certain dollar value or it can be dispositioned return
to drawing. But if you want to do anything else you go to a Material
Review Board, which is a board chaired by the NASA quality engineering,
includes the contractor engineering, contractor quality, and in the
case of ET we had a NASA engineering rep if it was a fracture-critical
part, which is over and above the minimum program requirement but
we felt was necessary.
This is documented on an MRB [material review board]. MRB can buy
basically an alternate design solution. If it’s a weld defect
they can buy a weld repair technique. If a hole is drilled oversize
they can buy a larger fastener. If parts don’t fit they can
cause the parts to be reworked. But essentially what they’re
doing is buying an alternate design solution, and this is the reason
on ET we insisted on having a government engineering rep, because
we did not want quality to be buying alternate designs. We did that
on Saturn, which I was never comfortable with.
We brought up from MAF every MRB. There were boxes and boxes and boxes
of them. Had them in the HOSC at Marshall, Huntsville Operations Support
Center, and sat down and had a team of Martin engineering, Martin
quality, NASA engineering, NASA quality go over every MRB to see whether
something in that MRB could have been a contributor to the Challenger
accident. It was apparent that it wasn’t. The foam on the tank
is good for short periods of time, maybe 15 Btu [British thermal unit]
per square foot per second. That’s called Qdot. The Qdot of
the gas coming off of the SRB was on the order of 700. It just cut
through the tank. There’s no way the tank could have begun to
survive that. Now if it had been a little bit further around the SRM,
where it wouldn’t have hit the tank, we could have probably
completed the mission before the SRB burned out. But as it burned
it widened the gap. When that gap hit the tank, the Qdot was just—plus
in addition to heating rate, it also was a highly aluminized propellant.
An aluminized propellant acts as a grit agent so there’s no
way the tank could survive.
“The impact of the Challenger accident.” Basically the
one change we made was—in the fracture toughness world we proof-test
every tank, we take it to 105 percent of design limit load. The theory
is that if we have the largest crack we can’t reliably find,
at this proof-test level the crack will burst. If it doesn’t
fail, then it says we’re good for eight flights. But during
proof test there are certain areas which we can’t fully get
up to the required proof stress. So for those areas, we do a postproof
NDE, non-destructive evaluation. We do a penetrant inspection. Say
if you get 80 percent, that stress level is enough to open up the
crack so that the penetrant solution will show it up. We had the philosophy
we’d only reinspect those areas that didn’t get fully
We pulled together a team of fracture experts, and made sure that
it was headed by somebody outside of NASA, a guy from Northrop [Corporation]
headed it up. We had Marshall representation, we had Johnson representation
and Langley and Lewis and people from industry. Two changes they made.
One is they recommended that all repairs be reinspected, regardless
of whether they were fully proofed or not. Which is I think a good
conservative position because most of your cracks are going to be
in repairs. We’ve never found cracks in parent welds that did
not have a repair.
The other thing is that Martin had put out a crack allowance handbook
where if you had cracks up to a certain size they were insignificant
and therefore they were allowing quality supervisors to buy them on
the floor. This board recommended that you not do that. They recommended
that you have an engineering review on all cracks, and that would
be the MRB. They said all cracks had to go to MRB. We made those changes.
It wasn’t a major impact, but it was changes in the right direction
I believe. That was basically the output from ET on Challenger. ET
was a victim on Challenger, not the aggressor.
I can ask, because it was a couple years later, the super lightweight
tank came on board. Would you like to talk about that?
I mentioned, prior to the time I became chief engineer, Marshall had
project engineering within the project, and we had a senior person,
and I had half a dozen engineers. Some of the other projects had 15
and 20, for the SSME particularly. I had by far the smallest staff.
I had some very very good people, some real self-starters.
At that time the chief engineers were in Science and Engineering,
and they reported to the deputy director of Science and Engineering,
and they were the interface with the lab experts. They had half a
dozen people on their staff who were specialists in the different
fields. When Jack Lee made the decision in ’88 to combine them,
as I mentioned, Jack Nichols, who was the chief engineer, moved over
to ASRM [advanced solid rocket motor], and I took over both roles
on ET, that is chief of Project Engineering and chief engineer ET
within S&E. My role within the project didn’t change but
I added the chief engineer role, and had to deal with Mr. [Robert
Mr. Schwinghamer is a most unique individual. He is a man of strong
beliefs. He’s a Purdue [University, West Lafayette, Indiana]
graduate. Excellent. He was director of Materials and Processes for
years, and then deputy director of Science and Engineering. One other
person that you may want to talk to is Paul [M.] Munafo, Dr. Paul
Munafo. He was with Boeing and Chrysler, and he is a fracture mechanics
expert. Within Marshall, the Structural Dynamics Lab did fracture
mechanics. Mr. Schwinghammer, when he had Materials Lab, decided he
wanted his own fracture mechanics expert so he hired Paul and brought
Paul in. Paul eventually was director of Materials and Processes and
got his doctorate on super light tank.
Interesting story. Paul’s undergraduate degree was MIT. When
he worked for Boeing he got his master’s at Tulane [University,
New Orleans]. Did his coursework for his doctorate, but Boeing would
not pay for him to go spend a year on campus, and he couldn’t
afford it otherwise. So when Marshall hired him he went through Auburn
[University, Auburn, Alabama] and got his doctorate, had to retake
all the courses because he lost everything that he had taken at Tulane.
When he was head of Materials Lab, because Auburn has a strong materials
program, a lot of his underlings were Auburn graduates.
Super lightweight tank. The ET was chugging along pretty quietly.
At one time Martin had owned an aluminum company. They’d sold
off the company but kept the labs. Let me step back a few years—in
the late ’50s Alcoa had developed an aluminum-lithium alloy.
This had been used on an aircraft called the A5J, which is an attack
airplane for the Navy. It was a Navy strategic bomber when the Navy
had a strategic role to deliver nuclear weapons. When the Navy’s
strategic role was taken away they converted these to reconnaissance
airplanes and they flew extensively during the Vietnam era. They were
Mach 2 airplanes. They were very good airplanes but they had a lot
of materials problems.
The material gave them a lot of problems so the industry pretty much
backed away from aluminum-lithium. The Russians had been using it
extensively but the US industry backed away from it. Then when the
composite world started coming along, the aluminum industry perked
up and said, “Hey we’ve got to compete with the composites.”
Aluminum-lithium is a lighter, stronger alloy so they started developing
aluminum-lithium. Alcoa developed one that is used on the [Boeing]
C-17 [aircraft]. But when we looked at it, their alloy was not particularly
weldable, and it tended to be laminar—that is if you made it
in thick sections it would be in layers. The strength through the
thickness, which is called the short transverse strength, was very
poor so we couldn’t use their alloy.
Martin in their lab developed what they call a family of Weldalites,
which is lightweight weldable aluminums. They came up with an alloy
that looked promising. They came to us and said, “Hey, we have
this alloy. We would like to deliver you an 8,000 pounds lighter tank.”
We went to Johnson, and Johnson says, “We don’t need performance.
We can launch now on a due east mission with more weight than we can
safely land on an RTLS [return to launch site].” So Johnson
said, “Don’t worry about it.”
At the end of the year we had $1 million or so left over. We said,
“Can we go buy some material and just get started? It could
be very useful for the next-generation program.” Johnson says,
“We can’t spend Shuttle money on next-generation.”
So we didn’t do it. Then our friends on the International Space
Station decided they wanted to have the capability for the Russians
to fly directly to Station so they changed the inclination from 28.5
degrees to 51.6. That cost Shuttle 13,500 pounds of payload.
You have to remember the history of Station. Station went through
about five design iterations, pretty much at congressional direction.
The last one passed by one vote. If Station had come back and said,
“We have to redesign again to a lighter configuration,”
I don’t think Congress would have bought it. We were already
off building the modules. Boeing was building the modules up here
at Marshall. The modules were already being built to the previous
weight. If we had had to go back and redesign it again, I don’t
think it would have survived.
Martin had been bidding against Boeing for the Space Station. Martin
was going to build them in New Orleans and Boeing was going to build
them somewhere else. We felt that that would open grounds for protest.
So our local office New Orleans people went in and figured out what
area of the factory could be cleared out, and we took Boeing down
there and said, “Hey, we can give you this space.” I was
part of the team who went down when we took Boeing to the selected
area. Of course Martin was very unhappy with that.
At every aisle Martin had people standing there to make sure the Boeing
guys didn’t depart from the team. We had NASA people escorting
Boeing. We took them, showed them the area, and walked back. One of
the Boeing people had been one of the managers on S-IC. His ex-secretary
was now the chief of the staff for the vice president of manufacturing
at Martin. They saw each other, went over, hugging and kissing. Martin
guy says, “I’m not supposed to let them talk—what
about hugging and kissing?”
wasn’t in the rules. That’s funny.
NASA folks were just laughing their heads off. But Boeing chose to
use some space here at Marshall. We had a Building 4755 that had been
built for Saturn and was in essence empty, and we turned that over
to Boeing and Boeing built the Space Station modules here on NASA
property, and that way we didn’t have to pay Boeing for space.
The Station was in a mode where if they went back to redesign there
was a good chance that Station would be killed so we had to come up
with performance. The Shuttle Program came back to ET and asked about
the 8000 pounds of additional payload we had offered. They convened
what’s called a Nonadvocate Review Board. It was chaired by
Bob [Robert D.] White from Johnson who was deputy chief of the Systems
Integration Office. It included John [A.] Wagner from Langley who’s
Langley’s aluminum-lithium expert, Tim [Timothy P.] Vaughn here
at Marshall who’s Marshall’s aluminum-lithium expert,
I think Neil Otte out of Marshall’s structural analysis, and
I was the chief engineer’s rep on the board. Martin Marietta
presented where they were. The aluminum-lithium experts said, “Yes,
we’re ready to come out of the lab.”
This was the material that Martin had in their lab, and they had sold
the production rights to Reynolds, but no production material had
ever been made. So they said, “It’s ready to come out
of the lab. What’s the schedule?” When Martin came into
an empty factory with no workforce and no tools, they delivered the
MPTA four years after they were turned on. So I said, “Four
years is probably a reasonable time.” That’s what Martin
had proposed. Johnson waited four months to turn us on, gave us 44
months to do the job. Since the agency had to have it ET felt that
was an acceptable challenge, so we got started.
Since Reynolds had the production rights, we went to Reynolds to cast
the first material. I was up there when they cast it, not that I know
anything about casting aluminum. In fact Mike [Michael E.] Lopez-Alegria
was with us. One of the technicians there was Hispanic and Mike went
over and started talking Spanish to him, and the guy had a grin wrapped
around his head three times. Overjoyed that an astronaut was talking
to him—Mike was in his blue suit of course. The first materials
that Reynolds cast didn’t have the same properties as the Martin
material. So we said, “You need to do a Taguchi design of experiments
to see what’s wrong.”
They said, “What’s that?” Martin had a young lady
here in Huntsville who, when we had to get rid of the Freon blowing
agent to go to another blowing agent, she had headed up the design
of experiments activities in the foam formulation lab here. She went
up and taught Reynolds how to do design of experiments. Their material
never was really consistent so every plate had to be tested. Every
heat from the oven—the size of our plate is 20 feet long, 12.5
feet wide and roughly two inches thick—every heat was one plate.
For every plate they would crop something off the end. They would
go in and they would EDM—Electrodeposition machining, which
gives you very, very fine notch—and they would pull it to failure.
Then they would take some more material, EDM a notch, and then stress
it to the level that it would be stressed in proof test, stress it
eight times at the level it would see during loading, stress it to
the level it’d see in flight, and do that four times. On the
last one we’d break it, and it had to break within a certain
percentage of the original notch. If it didn’t, that whole plate
went back in the oven and got remelted. We called it simulated service
test. We had to do a sim service test on every plate.
The first challenge is Johnson cut four months off the time. The second
challenge, we couldn’t make the material. When Martin took their
lab material out to Aircraft Hydroforming and when Aircraft Hydroforming
had tried to stretch it, the material was so stiff it broke the jaws
on the stretch press. The energy released look like an earthquake,
and it almost took the roof off the building. We had less time than
we needed, we couldn’t make it, and we couldn’t form it.
We tried to weld it. It didn’t weld like anything we’d
ever had. Had to have a back-side purge. Since most of our weld tools
cast the material into a chill bar there wasn’t room for a backside
purge. We had to redesign all of our weld tools. Normally we would
weld at four inches a minute. That was too much heat for this material,
it welded ten inches a minute. So we couldn’t weld it.
Martin had welded up three of these gores, and one of the problems
you always have when you make a repair, the repair tends to shrink
which tends to give you a flat spot. When you pressurize a compound
curvature dome with a flat spot, the flat spot pops out and it pops
the foam off. So Martin says, “Well, let’s practice our
contour repair techniques.” There are two or three different
techniques that you can use. They deliberately made a flat spot where
they made multiple repairs on the same spot. Lo and behold, it cracked.
Cracks in multiple repairs in aluminum is not unusual. But this cracked
so wide open, you could see light through it, read a newspaper through
it. So we couldn’t weld it, couldn’t form it, we couldn’t
repair it, we couldn’t make it, we had less time.
When we tried to figure out how to repair it, we brought in the Edison
Welding Institute, top welding people in the world. They looked at
what we had and said, “You’re never going to learn to
repair this material.”
We said, “Would change of weld wires help?”
They said, “No.” So Fred [P.] Bickley [III], who was working
on his doctorate on aluminum-lithium, was put in my office as czar
of weld repair. Fred still has their letter in his file. We had to
learn how to repair it. That gave us some interesting challenges,
and this is something.
One of our bright young stress analysts, guy named Pat [Patrick] Rogers.
Pat was a 4.0 graduate at Tulane. I’m a Tulane graduate, but
I was not a 4.0 graduate. When you make a repair and then you develop
the strength of the repair, you’ll make a weld, and then you’ll
make a repair, and you’ll cut a dog bone out of that repair,
and then you’ll pull it. You do a number of these to give you
a statistical database. That’s what you use as your allowable
for the repair in your structural analysis. Pat says, “We’re
cheating ourselves, because in the real world, when you load that
weld with a repair in it, the loads are going to redistribute. The
repair is going to stretch, and the loads will redistribute, and the
virgin weld will carry more of the load.”
So we said, “That sounds reasonable.” Pat had a finite
element model that would show in color the weld distribution. We had
a photo stress program that a visiting professor from the University
of Alabama [Tuscaloosa] had developed. A coating that you put on and
look at under UV. We put that on and we ran some 17-inch-wide panels.
Started off on 2219, which was the old alloy that we thought we understood,
and it did exactly what Pat was predicting. We got much higher strengths
than we were taking credit for. Then we said, “Let’s try
that on aluminum-lithium.”
We were shooting for about 32 ksi [kilo-pound-force per square inch].
It failed at about 17 or 18. What was happening, when you make the
repair you get a certain amount of shrinkage, and the shrinkage was
subtracting from the capability. The material was so stiff, the loads
didn’t redistribute. So they said, “What can we do?”
Well, let’s get rid of the shrinkage. Before you make the repair,
you have to scribe two lines of it, measure the amount of shrinkage,
and then go and planish. Planishing is where you use a rivet gun with
a mushroom head, and a guy on the other side of the bucking bar, and
force the weld bead back into the metal, and spread the metal back
out to get rid of the shrinkage. Well, if you’ve got a 27.5-foot
tank and you’ve got one guy on the outside with the rivet gun
and a guy on the inside with the bucking bar and they can’t
see each other, and they’re communicating through a headset,
and if they get out of line, you’ll be putting a moment into
the weld. It’s challenging.
One of the requirements we put in is that before you make any unique
kind of repair you have to verify it on two 17-inch-wide panels. At
the first R17 repair where we had made 17 repairs at one spot, to
make the first test panel took 10,000 hours. The material is very
unforgiving. It is not a user-friendly material. With friction stir
welding it’s a different animal. Friction is beautiful, but
this was fusion welding.
But we were able to develop it, were able to deliver the 7,500 pounds
lighter tank. The initial tank weighed 78,000 pounds. The tank that’s
flying today is 58,000. So that’s 20,000 pounds of payload.
The payload to Station is about 35,000; 20,000 of that came from weight
reduction on the tank. The rest of the program, orbiter did some,
but not a whole lot. The orbiter is so complex that they just weren’t
able to do much.
When you get into the Columbia accident—I had retired five years
before Columbia. I really don’t think I should be talking about
the Columbia accident. The people you can talk to is Steve [Steven
G.] Holmes—if you’re looking at NASA people. John Chapman
was project manager in the return to flight. John has now retired.
Steve Holmes is currently NASA, and he was the materials lead within
the project office. Scotty [J. Scott] Sparks was our materials guy
in the lab, and then he moved over to the project. He’s deputy
chief engineer. Ken Welzyn’s is chief engineer. These guys can
talk in depth what went on in the Columbia investigation and return
to flight, what they’ve had to do. They have done a fantastic
job, but as I say, since I had retired five years before—USA
brought me back as a consultant during the investigation, but as a
consultant you’re on the outside. I was not in a lot of the
NASA-Boeing-Martin meetings because as a USA consultant I wasn’t
invited. I didn’t feel that I should just go charging into meetings.
makes sense. You covered the Challenger very well. Is there anything
else that you can think of adding right now about your experiences
with the tank that we didn’t cover? We talked about digital
X-rays—did you introduce those at some point?
typically we previously had used film X-rays. That gets expensive,
because the film is quite expensive. We did go into digital X-ray
in some applications where you could put the digital receiver in.
Of course this lets you expand and do a lot of work with the pictures.
For the foam they went in and started doing a lot of work trying to
analyze the foam capability.
One of the techniques that Scotty or Welzyn can talk about—you
fire an X-ray beam in, but it hits the aluminum surface and reflects
back. You’re looking at the attenuation in the beam as it reflects
back. This was to try and find voids in the foam. If you fire something
energetic enough through the aluminum substrate, it just washes out,
you don’t see anything in the foam. There were two X-ray techniques
that they have tried to detect foam. One is not exactly an X-ray,
it’s an ultrahigh-frequency.
Foam is such a good energy absorber that it’s very difficult
to get any meaningful data, but they do have those techniques. As
I say, last time I was familiar with them—Scotty can tell you
more—it was a rastering process. That means it’s very
very slow. When we’re looking to do this on a production basis,
you have to have the part in a location where you can do it, and you
tie up—again, when you’re trying to produce parts, you
can’t park something for two weeks while you scan it. You got
parts that have to flow. I won’t say you can’t—you
could, but it’s not conducive to production flow so that’s
Really it’s also workstations. You’ve got a limited number
of workstations, and if you tie up a workstation for long periods
of time, then everything else backs up. I know we were moving along
pretty good—the initial cost of an ET was about 400,000 man-hours,
about 200,000 of that on TPS [thermal protection system] and 200,000
in metal parts. At the time of Columbia I think we were down to 200,000
with 100,000 and 100,000. The man-hours had come way down. Now after
Columbia the TPS cost just soared astronomically.
curious. I know there was an expectation to do 60 a year. How many
at one time did you have?
the time of Challenger we were building at about 17 I believe.
had backed off from the 60 to 24. We had done what we called a 60
minus 36. We had facilitized the plant for 60—whether we would
have ever been able to get there is another question—but we
put in tooling for 24. We left the open workstations so that if we
needed to go up to more we had space to install more tools, but the
tooling was supposed to support 24. As I say, at the time of Challenger
we had ramped up to about I think 18. Again we didn’t have the
extreme TPS activities that we went into after Columbia, so we probably
would have been able to meet 24.
curious how many of them are left, how many tanks?
There are two of them set for the next two flights, and there’s
one that was in stacking position in the Vertical Assembly Building
at Michoud during Hurricane Katrina, and some of the roof fell in
and some of the concrete fell and hit the tank. They have gone back
and repaired it since then and I believe it’s been recertified
for flight. But the way the program is set up, when we fly we’ve
got to have a rescue vehicle on the pad, and that tank is the tank
for the rescue vehicle. Now whether the program is willing to fly
it without a rescue vehicle—they initially said, “We’re
flying to Station, we can park at Station.” Then they backed
off on that. You’ve got to have a rescue vehicle in case there’s
a problem before you get to Station. Of course you had to have that
when you went to the Hubble [Space Telescope] because you couldn’t
really park at Hubble.
room at that inn, is there? Based on all the improvements and enhancements
up until now, how long would it take from beginning to end to put
a tank in place?
can’t say. I know in the welding area there’s one tool
that has been removed—we’ve been looking at this under
the heavy-lift vehicle. If we go to a 27.5-foot heavy-lift, could
we build a 27.5-foot heavy-lift?
Martin has done some studies on that, Marshall manufacturing engineering
guys have done some studies, and the ET project. Basically from the
standpoint of building the tankage, there’s one tool that was
removed to clear out areas for the Ares I upper stage. We think there’s
a work-around on that tool. There’s a robotic friction stir
weld tool that National Council for Advanced Manufacturing has down
in New Orleans. We have been partnering with NCAM developing new manufacturing
techniques. The state of Louisiana in essence built a tool—the
state of Louisiana is partnering with them also. For the heavy-lift
we have to build a thrust structure and other stuff, but as far as
building the tanks, the tooling is basically there now as far as Martin
is concerned. NASA put out a request for information, and Martin came
back with a study. If you want to talk to Martin, the local Martin
manager is a fellow name of Ron [Ronald W.] Wetmore.
Ron is an unusual individual. He was in the Naval Academy, and at
the end of his junior year Admiral [Hyman G.] Rickover came to him
and said, “If you get your grades up you can get into nuclear
school.” Ron said, “I don’t want to go to nuclear
school, I want to fly F-14s.” Well, when he graduated, he got
his grades close to that level, and he got assigned to nuclear school.
I’ve later learned that there was one or two years there where
Rickover didn’t get as many candidates as he wanted, so he was
drafting academy graduates to nuclear school. Ron was chief engineer
on a nuclear ballistic sub and on an attack sub, was in line to be
a captain of a nuclear sub.
He had a family and kids, so he got out of the Navy and Martin hired
him as their chief engineer down at the Cape for a number of years.
Then they brought him back to Michoud, and he’s now heading
up the Martin office here that has been bidding on stuff and he has
two master’s degrees. He’s just an outstanding individual,
with an unusual background. Very very nice person. Ron is an extremely
thank you so much for being so kind with us today and offering so
much of your time. I really appreciate all the information. It’s
been very valuable.