Accident Investigation Board Roundtable
June 24, 2003
429 L' Enfant Plaza, SW
MS. LAURA BROWN: Okay. Thanks for waiting again, and
I think most of you know Scott Hubbard. He's been in charge
of the foam testing done in San Antonio, and he's going to
give us an update.
MR. SCOTT HUBBARD: Okay, good afternoon. This is part
of an ongoing series of tests of impacting the foam against
the wing leading edge in order to establish the connection
from foam to breach. The impact on the RCC panel six, I think,
gave us a plausible failure scenario, and what we're trying
to do here is to see, by doing these impact tests, can we
establish that we have the data, as Admiral Gehman said a
little earlier, to say the words "highly probable"
or "most likely." So, that is the purpose of this
of these tests.
Okay. So, let's go to the first slide. We have conducted,
since we last talked to you, two fiberglass tests, one that
was on June 16th and one that was on June 18th, and these
were designed to answer three questions. I think I told you
about this the last time we were together. But, just to refresh
your memory, first of all, we wanted to understand do these
panels act as a system, or are we simply looking at the interaction
of the foam hitting a single panel or a single T-seal. And
I have a new model here, which I'll refer to several times,
that shows how the test article has been structured. This
is the piece of material that we're using down at the tests
in San Antonio, and it goes from panels five, six, seven,
eight, nine and 10, and what we're trying to find out, first
of all, is when you hit this with a piece of foam, does the
entire wing-leaning edge respond, or are we looking at a localized
phenomenon that created the breach, and I'll come back to
this as we talk about the results.
Second thing I want to talk about is the question of where
the foam hit. The visual record, the sensor record, doesn't
tell us that exactly. You've heard from other board members
that the interpretation of the debris evidence and some of
the sensor evidence seems to indicate that the breach was
in the lower part of panel eight, which is this panel here,
but we don't have, as yet, any experimental evidence that
establishes that the foam, in fact, could create a breach
of sufficient size to create the accident. So, what we're
doing is removing the impact position of the foam down closer
to where the debris evidence indicates that it probably occurred.
That's part of the set of fiberglass tests, and will be part
of the next and final series.
Final thing is what happens as the foam impacts the wing-leading
edge with different areas? Couple of you who have been to
San Antonio have heard me speak of the term "Clocking Angle."
I'll show you in a picture clearly what we mean by that. But,
the point of the test is, does it make a difference. As this
piece of debris was tumbling and spinning, whether or not
it hit just with a corner, or whether it hit with an entire
edge against the panel, and how is the force distributed.
These are all things that we're trying to take into account
to be sure that we have the most representative footprint
of the foam as it hits the panel. So, those are the questions
we're trying to answer by these last two tests, the system,
what happens when you hit lower on the panel; and finally,
what happens when you apply the force along the entire edge
of a piece of foam rather than just hitting on a corner. So,
let's go to the next slide.
This is our setup in San Antonio. Many of you have been there
in person with our nitrogen pressurized gas gun, 30-foot barrel.
We see all the lights and the high-speed cameras, and you
see the wing-leading edge, panels five through 10, which has
been duplicated in this model here. Lot of data-taking, several
hundred channels of sensors, as well as about a dozen high-speed
cameras. Now, the next slide.
The two tests were designed to answer the questions that I
described. The first test, we had a target that was about
three inches lower than it had been on panel six. As you know,
we are using aiming for consistently a velocity of
775 feet per second. That's a consensus on how fast this was
traveling when it fell off of the bipod ramp. In this test,
we were able to achieve 774. All of the tests thus far have
been well within a percent or so. So, the accuracy of the
gun, both in velocity and in aiming, have been quite good.
Now, we did an additional thing in here to look for the systems
effects. We instrumented panel eight. If you recall, here's
panel five. Here's panel six, where we're aiming, panel seven,
where we've evaluated the response thus far. And we instrumented
panel eight to see if this impact travels downstream very
far or simply terminates within a panel or so. And I'll come
back to the results of that in a minute.
Standard foam size, the 1.67 pounds that we've been using
all along. The second the next test, on June the 18th,
actually is called fiberglass test three because we conducted
one several weeks ago, was the same as the test on the 16th
with one exception and, if you look at the picture being projected
up there, you can see what I hope indicates the clocking angle,
that is to say the barrel has been turned by 30 degrees. Now,
what does that do for you? What it does and I'll take
off one of these RCC panels here what it does is that,
by rotating the barrel 30 degrees, rather than hitting the
panel on the corner first, you actually hit the entire edge,
this 11-1/2 inch edge hits on the panel, and the analysts
predicted this would, in fact, transfer more force locally
and would, therefore, concentrate the energy to make it more
like what we believe happened during the accident. That was
the purpose of the second test.
So now, let's talk a little bit about the what actually
happened during the test. The first video, then. This shows
test number two, and you can see the streaks from the previous
test. We left them there. You can see the panel the
foam impacting it, and see it hitting the T-seal and creating
a cloud of debris. I think some of you have seen other versions
from further away. I decided to show you this so that you
would get a sense up close of what it looks like when the
impact occurs. This loops over, and you can see that it hit
on the corner first, and then if you go to the next slide,
you will see how it actually impacted. And if you can see
here, right there you see where it actually hit in this test
versus the previous test. This is the difference. This is
the three inches or so lower down. But, note that the impact
footprint here is triangular. It did spill over into panels
seven and even eight, to some degree.
The interesting thing about this test was that the step, that
is, the difference between where the panel was and where the
T-seal was, was about six-tenths of an inch, almost five-eighths,
rather than less than half an inch for the first fiberglass
test. So, more force was being transferred in this local region,
consistent with the prediction. And you can see here, too,
how much the foam got stuffed under the T-seal, creating a
as we noted some time ago, perhaps an additional failure
mechanism whereby the T-seal gets shoved aside by the impact
of the foam. We did see one small crack in the fiberglass,
about an inch long, and as you see the results of the third
test, you'll see how the force was even greater.
So, let's go to the sensor results from test two, and remember
what test two was all about was firing lower down in the panel,
seeing if the force down in the region where the debris indicates
the breach occurred was, in fact, higher. Then and
whether or not this system of panels responded together or
responded individually. And what we found was that we had
very high stresses down here on panel eight. So, this whole
leading edge is responding as a system. We hit here, several
feet away, and stresses creating a measurement of 60,000 pounds
per square inch occurred down here in panel eight, indicating
that this impact had been transferred down several feet, and
so the whole thing is, in fact, responding together. We can't
simply say that one panel or one T-seal constitutes a test,
and this is important later on as I describe the what
we're going to do next.
The face sheet motion, that is to say how much this went in
and out, how much this particular area here called a face
sheet went in and out, is a little over two inches compared
to an inch and a half for the test on panel six. So, that
showed us that there was greater local force in the bottom
part of this panel, and the sensors all showed also
showed higher stress, as well. So, we learned quite a bit
in doing that test. Going on to test number three, it's the
same setup but, again, we rotated the barrel so that we could
place this entire edge here of 11-1/2 inches flat against
the panel. The prediction was this would actually result in
more local stress, although maybe lower overall stress.
So, let's look at two videos there. The first one shows, again,
a close-up impact of the foam on the face sheet, and you see
now the entire face hits it. It breaks up. If you were to
look at it from further back, you would see the entire piece
of debris disintegrating. And the key thing that you'll note
here, and that you'll see later on as well, is that the footprint
is different. It is a more square type footprint as opposed
to the triangular shaped footprint, which you might expect,
given that you impact on an edge here rather than on a corner.
But, it tends to be important in our thinking about how this
breach may have occurred, how you were able to put so much
force in an area in panel eight and actually create the hole
that most of us believe is there, or knock a T-seal aside.
Let's go to the second video from test number three, and what
you're going to see here is the inside. This is one of our
six interior cameras, and I think you're going to get a sense
of the violence of this force of this impact
as it deflects the entire panel. And you're going to see some
brightening. You see the foam coming through, and you'll see
little flashes of light, and that's actually the panel being
spread far enough apart to let daylight in, and then it closes
up again. This was not predicted in advance, but it shows
you how much the interaction between the panels and the T-seals
and all the whole system is going on.
Okay, let's go to the results from this test. Here's a couple
of visual results. We now have, as I said, a footprint. If
you see the streaks down here, it is a squarish area. I think
we're beginning to close in on the kind of impact that probably
occurred that created the breach that led to the destruction
of the Orbiter. The T-seal is pushed aside. There was a lot
of foam in there. But, an unanticipated result was that four
large cracks appeared in the fiberglass. And fiberglass, if
you remember from the previous briefings, is something like
two times or more tougher in certain measurements than the
reinforced carbon panels on the actual Orbiter.
If you go to the next slide, you'll see the largest single
crack, which was over eight inches long. It is this rather
ugly looking thing right here. So, the next slide shows the
summary of the sensor results from this test. We saw four
cracks that had not been previously observed. Since this panel,
panel six now, had been hit twice before, one can naturally
raise questions about of whether or not this is a cumulative
effect or not. In looking back at the evaluations that were
done, there was no crack except for the one right after the
second test. Visible, certainly nothing of the dimensions
of what we see here. So, the sense of the test group is that
these four cracks were created as a result of this test, indicating
that a lot of force was transferred.
The face sheet motion, that is to say how much this got moved
in and out, and you saw form the interior camera how violent
that was. That total excursion, that total distance, was more
than three inches, much greater than the previous tests. And
finally, we measured up to 50 percent more stress in panel
six, as well. So, the conclusion, then, is that this angle
of impact does make a difference, and that it transfers far
more force, as predicted, to the impacted area.
Overall then, going to the next slide, where are we? Well,
by measuring panels six, seven and eight, we believe now that
we have established that this leading edge system actually
acts as a system, and since our models have been pretty good
on predicting stress but not good on predicting breakage,
the best way to simulate what happened during the accident
is to make the final test on the reinforced carbon as close
as possible to what we believe happened. That means you've
got to test at least three panels in a row. In order to do
that, it means that we're going to shoot at panel eight, not
panel nine as we thought a few weeks ago. This is what
a major thing we learned from the test, so that we've got
two panels downstream from it so that we simulate this system
The debris evidence, as you heard several of the board members
talk about, indicates that the lower part of panel eight was
probably the likely site of the breach. The forces and the
footprint that seem to most closely correspond to creating
that sort of breach, were the conditions we used in fiberglass
test three. That was the point of doing the fiberglass tests,
not that fiberglass is exactly like RCC in its detail properties,
but that by comparing different conditions, we can establish,
hone in on, box in the kind of conditions that would be, in
our thinking, the kind that would create of some substantial
size in the bottom part of panel eight. Therefore, the final
RCC test that we're going to do should be in the lower part
of panel eight, the same setup that we used in the third fiberglass
So, final slide, "What are our next steps?" We're developing
a test plan together with the folks at Southwest Research
Institute and the people from all over NASA that have been
participating, and it'll be a test using all reinforced carbon
panels and T-seals from eight, nine and 10. We're going to
take in fact, it's in the process right now
of panel eight from Orbiter 104, which is Atlantis. It has
27 flights on it, so it will be very close to the type of
flight history that Columbia had. Just as a note, there's
a new panel eight on order since this was the one spare in
the fleet, and delivery time there is probably something,
six months or more, but that is probably matches up
reasonably well with the return to flight schedule.
We're going to test this panel eight per what flight specifications,
the same process that panel would go through if it is being
used and created for the first time by the vendor. These test
here, visual, x-ray, eddy current and so forth. If there's
time to fit it in, we'll do this so-called thermography test.
And the schedule is the following: we're going to test Friday,
a shot at fiberglass panel eight. That's the 27th. If necessary,
we may do another one on June the 30th to be sure that we've
got the positioning and the footprint and the velocity just
right. And the schedule right now for doing the all RCC test
is July the 7th. We believe this is a pretty firm date. Unless
there's a big surprise, all of the testing and preparation
and so forth should line up to where we can do this very important
test on July the 7th and, hopefully, add to our knowledge
about the most probable cause of this accident.
That concludes my remarks. I'd be happy to take questions.
MS. BROWN: Okay. I think we're going to try to take
a couple questions from the phone bridge first, just because
we're going to lose the bridge in a few minutes. Bill, are
you still there?
UNIDENTIFIED MAN: Yes, I am.
MS. BROWN: Okay. Do you have a question?
UNIDENTIFIED MAN: I do. Scott, just a quick one for
me. Since you're taking panel eight off of 104, are you going
to do any NDE testing to kind of get the before and after,
if you will? And if not, I mean, isn't that an important thing
MR. HUBBARD: Right, yes. You probably don't have the
slides there with you. I believe they're being posted to the
Web site. But, there is a on the last chart there,
that we'll do non-destructive evaluation just as the
at the vendor site, just like they would do for a brand new
panel, to establish a baseline, and that includes a visual
inspection, an X-ray, what they call eddy current, which is
a type of electrical test, as well as ultrasound. And even
though it's not done as a normal vendor inspection, if we
can work in a thermography test that's an infrared
test we'll do that, as well, so that we've got a good
MS. BROWN: Gina, are you still on? Okay.
MR. PHIL CHEN: Phil Chen's here.
MS. BROWN: Okay, Phil, go ahead.
MR. CHEN: Okay. Scott, early on, there was talk about
aging in the RCC panels being a possible cause. Admiral Gehman's
example with the termites is that still being addressed?
I'm thinking are you thinking about acquiring a virgin
RCC panel to see how different it is from one which has had
27 flights on it?
MR. HUBBARD: Well, what we're doing in this test is
trying to establish the most probable cause of the initiating
event and trying to connect the dots between the foam shedding
and the hole that was almost certainly in, say, the bottom
of panel eight or nearby. And so, we're focusing on trying
to create the flight conditions as closely as possible, and
to us that means using aged RCC, the same flight history,
if possible, as Columbia. I don't have any doubt that NASA
and the shuttle program in the future will probably carry
out a very extensive series of tests of material of different
ages. I know that there are is work going on already
to look very carefully at the kind of sub-surface defects
that would occur as part of aging. But, I believe that's future
work. What we're focusing on is trying to re-create the conditions
of the accident.
MS. BROWN: Anybody else on the phone bridge?
MR. KEVIN SPEAR: Laura, it's Kevin Spear.
MS. BROWN: Okay, Kevin, go ahead quick.
MR. SPEAR: I'm wondering, what do the models say about
the transfer of energy in an all RCC system versus an all
fiberglass system, the rate and the amount of transfer of
MR. HUBBARD: The models have been pretty good in predicting
strain, that is to say how much flexure and bending and tension
are these under, and they have not been particularly good
at predicting breakage. We have a case where something broke
in the test on RCC panel six just above the so-called allowables
that's what the manufacturer provides. But, another
place in the panel, it was three times the specification and
didn't break. That's why we are making this test as much like
the real wing-leading edge as possible. The difference between
tests, I think, is an important way to be able to translate
from fiberglass to RCC. We can't, in detail, make the connection,
but the fact that one set of panels shows this transfer of
energy, that the models show that the stress accumulates in
a similar way, I think gives us confidence that we're doing
the right thing in putting an all-RCC panel together.
MS. BROWN: Anybody else on the phone bridge? Okay.
Sorry if we lose you guys, but I believe you can watch the
Web cast version to follow along. Why don't we do questions
the same way we did before? Todd?
MR. TODD HALVORSON: I'm Todd Halvorson of Florida Today.
Interpretation in the debris in the sensor data in evidence
now seems to come to the conclusion Mr. Tetrault said
that foam hit was the most likely cause of the accident, or
that it was highly probable that foam cause was initiating
the event. In light of your foam testing to date, where do
you stand on that?
MR. HUBBARD: I think that we have I agree completely
with Roger, that, based on the data from the debris and the
sensors and the way the wires burned through and so forth,
that there had to be a breach probably somewhere in the bottom
of panel eight, maybe in the so-called T-seals, you know,
these things here that are either side of a pane. And we also
know without a doubt that a piece of foam that weighed about
less than two pounds came off and hit at about 500
miles an hour. The thing that I'm trying to do with these
tests, and I think the board is looking at this information,
is a piece of that most probable cause is to connect that
dot from foam to breach. And so, that's where I stand.
I think that the panel six test that we did showed that we
got a plausible failure scenario. We created a substantial
crack five inches long, but we haven't created yet a breach
that is like what has been described by the debris. So, you
know, maybe I'm one step behind Roger in coming to a conclusion
but, as Admiral Gehman said, that's part of our discussion
here is to put all the data on the table and see if each and
every one of us agrees on that adjective of most likely, highly
probable, so forth.
MS. BROWN: Mark?
MR. MARK CARREAU: Mark Carreau, Houston Chronicle,
thanks. So, what do you predict is going to happen on July
the 7th to connect the dots? What do you think you need to
see? And I'm I guess I'm talking not only about panel
eight itself, but if you're going to measure outboard of that,
what sorts of things do you think you need to see to connect
MR. HUBBARD: Based on the tests on panel six, you know,
where this piece of information what it adds to the
total discussion about plausible to most probable, I think
will depend on the extensiveness of the damage. You know,
is we see a whole network of cracks? You know, do we
see something that runs across the T-seal? You know, given
that our models have been pretty good in forces but not so
good in breakage, I'm not going to go out and say what I would
predict there. I'd rather let the system respond to us and
tell us, but I think that, you know, from a crack to multiple
cracks, or maybe fractures or breakage, will help us set the
adjectives from, you know, plausible to highly probable.
MR. CARREAU: It doesn't sound like you're saying there
has to be a hole there.
MR. HUBBARD: I don't think that there has to be a hole
as a result of the initial hit. Remember, this occurred 82
seconds into a flight that had several minutes to go, and
had to go through more heating, more vibrations, sixteen days
on ordinance and so forth. And, you know, how these fractures
propagate and so forth is something that needs to be understood,
and it is not understood today.
MS. BROWN: Marcia?
MS. MARCIA DUNN: Marcia Dunn, Associated Press. This
is coming up pretty close to with the reports coming
out, and I'm wondering if are you under the gun to
get this out of the way before the final report is out? And
if you can't do it, for whatever reason or it's an incomplete
test, are you prepared to go with all the tests to date, minus
the last one, and put that into the final report for the conclusion?
MR. HUBBARD: I think that what it's a very good
question, and we don't want to rush an important test. You
know, we're taking our time to be sure that it's taken
months to get to this point, to build the right structure,
to instrument it properly. We added in this instrumentation
on panel eight that we hadn't added thought of before
because everyone was focused on the panel or the T-seal. And
what we started to realize was it's the whole system. And
so, we've tried to be careful about doing a good test to establish
whether this was the most probable initiating event. We haven't
made a science project out of this in the sense of understanding
all of the properties of reinforced carbon. That will come
What if, for some reason, we came to the point where
the test simply couldn't be done, or whatever, then I think
that the board would look at all the data on the table and
would make come to some consensus about whether the
this was the most probably initiating event or not.
I think, as Admiral Gehman has pointed out several times,
you know, this proximate cause discussion, this direct mechanical
cause, is only one part of a much larger discussion about
the shuttle program. At this point, I think that the date
of July the 7th, in terms of conducting the test, seems to
be pretty good.
MS. BROWN: Kathy?
MS. KATHY SAWYER: Kathy Sawyer, the Washington Post.
Scott, do you you talked about the you showed
us the foam stuck in there, and you said earlier, I believe,
that this had not figured largely in your predictions until
you saw it stick in the first test.
MR. HUBBARD: Right.
MS. SAWYER: Is somebody else working on the possible
impact not real impact, but the possible effects of
the foam remaining stuck during launch, during arrival
in orbit, during maneuvering, etc., and is that playing, also,
into other these studies they talked about earlier
today about the launch properties of the launch?
MR. HUBBARD: Yeah, that's a good point. Up until we
had done the first set of tests and seen this mechanism that,
I mean, seems obvious in hindsight, of course, you know, if
you punch in a panel that's going to get caught underneath,
but this is why you do the experiments. It's only obvious
in hindsight. As soon as we did that, then the people who
were thinking about how the heat got in and propagated through
the wing began to say, "Ah, aha, you know, this makes
this is a different way to create a slip." You don't have
to have an entire T-seal missing. You stuff this foam underneath
it. You wedge something aside. Maybe the pins underneath get
permanently deformed and I'm hypothesizing here
but it's another mechanism to create a slit to let in the
heat. So, from that perspective, I think the test has told
us something that we were not aware of in the beginning, or
hadn't thought about.
MS. SAWYER: (Inaudible).
MR. HUBBARD: Yes. There are aerothermal the
analysts that do the thermodynamics and the aerodynamics are
looking at slits and how big would one have to be, the total
area and so forth, in order to let enough heat in to create
the kind of damage and the time sequence that had been observed
MS. BROWN: Ralph?
MR. RALPH VARTABEDIAN: Ralph Vartabedian, L.A. Times.
In past weeks, there's been a lot of discussion about panels
five, six and seven, and particularly the T-seals between
panels five and six as the possible breach point. And I recall
Dr. Widnall had described whether half a T-seal could be missing
and whether it would replicate the pattern of the mystery
object re-entering the atmosphere. And at some point, I believe
that's why, if I'm not mistaken, although I must be, that's
why you picked panel five for your initial or panel
six for your initial RCC tests. At what point did you begin
to go from panel six to panel eight as the most probable breach,
or am I missing something?
MR. HUBBARD: No, you've got it exactly right. What
you've experienced with the board is, you know, a certain
kind of physics in action, I mean, research and development
on the fly. We have taken the data, the best data of the shuttle
program and that NASA's come up with. They've refined it.
The tests have been refined in response to that. If you recall,
the very first tests we did were against the main landing
gear door, and that was because some of the early indications
were that was where the foam hit. As the visual analysis of
the film and the video got better, got cleaned up and de-blurred,
as the physics trajectory analysis got added to the
and married with the visual analysis, it became clear that
the footprint probably was not across if you remember
some of those early pictures, it was something like this coming
across here, and it actually moved the line of most
probable impact moved over here.
At the same time, the debris evidence began to come in. We
were finding some things and not finding others, and the timeline
of the sensors as they burned, the wires burned through and
so forth, began to come in, as well. So, all of that led us
to the point of testing, you know, panel eight as the most
probable, and panel six has turned out to be a good lead-in
to that to understand all the other system effects.
MR. VARTABEDIAN: Is this the first I know that
panel eight was always looming up. There was a lot of evidence.
Is this has there been some sort of shift just in the
last two weeks from panel six to panel eight?
MR. HUBBARD: No, no. We've had this this shift
occurred maybe, you know, a month and a half or two months
ago, yeah, something like that. But, the and in fact,
at the point where the data was starting to shift from panel
six to panel eight in the analysis of all of the trajectory
data, we weren't sure what the right thing was to do in building
up this test article. So, after a lot of discussion, I said
we ought to plan for maximum flexibility, and that's why this
whole thing got constructed, was because six, eight
you know, six to nine, so NASA got 600 pieces and, out of
seven different centers, put together so we could have some
significant flexibility to test the whole range.
MS. BROWN: Frank?
MR. FRANK MORRING: Frank Morring, Aviation Week. Scott,
will this data be used beyond determining the cause, or probable
cause, of the accident? For example, in recommendations that
the panel might make in hardening leading edges or other areas?
MR. HUBBARD: Yes. The data certainly is going to be
used for the future, and I think that we've already, as a
result of doing this investigation, caused some new data to
be generated about the properties of reinforced carbon and
all this we've got hundreds of sensor measurements
that are we've analyzed only the most important ones.
That'll be part of the database for the future. And one of
the things that we're looking at, thinking about in the preliminary
or interim recommendation category, is whether it's possible
to establish thresholds, you know, that so much energy represents
a certain type of damage and concern but, above this energy,
it represents a very serious concern. So, that data is playing
into those considerations.
MS. BROWN: Traci?
MS. TRACI WATSON: Traci Watson with USA Today. Along
the lines of Frank's question, have you considered whether
it's possible to armor the leading edge? You know, NASA's
always assumed this area is tough, but is it even feasible
to make it tougher?
MR. HUBBARD: It is. That's a good question for the
Materials Science people. The reason that this is reinforced
carbon is because that was, at the time, the best choice of
material that you could obtain that could withstand
and I have to qualify withstand. It's a little complicated
but, let's just say withstand 3,000 degrees Fahrenheit.
It's hard to find materials that are both tough, workable
and withstand those kinds of temperatures, and this carbon
material, when it was being designed in the late '70's, was
the best available. I think people are going to go back now
and think about whether, for the next vehicle, some other
material might make a better leading edge. I think for the
Orbiter, as it stands, you know, they need to work with what
they have and, as you heard, we're thinking about some recommendations
about how you might do an emergency repair.
MS. BROWN: Paul?
MR. PAUL RECER: You're going to do this on July the
7th in San Antonio in temperatures of 100, 105 are
not unusual in that area. And is this thermal difference ?
MR. HUBBARD: Are you going to be uncomfortable
if you come? Is that no.
MR. RECER: I've lived there. I love San Antonio. But,
if the thermal conditions are going to be different at the
test site than they were near 81 seconds into launch, obviously,
are is this any concern that this may bias the results
when trying to draw any conclusions?
MR. HUBBARD: Good question. It could have been part
of our test team. We asked those questions. We did a series
of tests that were done at Glenn Research Center to look at
cold, ambient, vacuum, regular atmosphere, sea level, and
so forth, and we also looked at the properties of the foam.
And this foam has the withstands on a normal flight
-300 degrees on one side and +400 300 degrees on the
other side. We looked at the temperatures in the barrel. We
looked at the thermal coefficients, you know, the various
ways in which the foam can respond, as well as the reinforced
carbon, and there nothing jumped out at us as being
anything that would bias the test. We did think about that.
In fact, the first couple of tests we did when it was 98 degrees
in the shade, you know, I raised that very question, and none
of the data we have has indicated that it's going to create
a false answer.
MS. BROWN: Okay. Earl?
MR. EARL LANE: Earl Lane with Newsday. Given the kind
of face motion that you saw in that test number three, I mean,
what can you predict might happen with that RCC panel number
MR. HUBBARD: Direct comparisons between the detail
properties of fiberglass and the detail properties of carbon
reinforced carbon are hard to do. That's why I emphasize
the difference between the tasks as being the more important
thing. Reinforced carbon is stiffer, but not as strong, as
to say it's more brittle. So, I can't predict exactly how
much it would flex. What I can say is that, comparing fiberglass
panel six in that first test and fiberglass and reinforced
carbon test in its first test, it was a half an inch of motion
versus 1-1/2 inches of motion. Now, is it a factor of three?
I don't know. You know, I mean, that is why we're going to
do the experiment. But, I think if the comparison holds true,
we would expect a lot more flexure, maybe even to the breaking
point, of the reinforced carbon, based on the comparisons,
the relative differences in the fiberglass.
MS. BROWN: Alan?
MR. ALAN LEVIN: Alan Levin with USA Today. Scott, could
you I can't quite tell from the photo here where the
cracks were. I'm assuming they're on panel six, but if you
could take us through where they were on your model. Also,
did you do any non-destructive tests in between the tests
so it was to give you a sense of whether there was
any un-visible internal damage?
MR. HUBBARD: We have done a fair amount of non-destructive
evaluation on the reinforced carbon panel. We have done some,
but less, on the fiberglass, because what we were looking
there for was the total response and how the one test compared
to the next one. But, just to take one of these panels here,
if you look at the hardware, and this is says fiberglass
test three results continued I don't know if you can
back up to that one or not. But, it is the lock side that
is you know, there's two sides to this, what they call
the slip side and the lock side. The lock side has the groove,
and it's inboard inboard lock side, and the
it was up in here along this rib that you had the eight-inch
crack. And let me be sure I'm saying that the yeah,
it's from the lower flange, so, remember, we aimed at the
bottom part of the panel. So, it's down in here.
MS. BROWN: Okay. Is that okay. Matt?
MR. MATTHEW WALD: Scott, Matt Wald with the New York
Times. The clocking angle on the actual accident, which is,
of course, unknown and unknowable, does this raise the possibility
that, whether the thing broke enough to produce a breach in
the accident, depending on the randomness of the angle at
which this tumbling block struck the leading edge?
MR. HUBBARD: Well, I don't know I can't give
you an exact answer to that. All we can do is speculate. We
do know from the visual evidence that this piece of foam was
rotating at about 18 times a second, and we have because
that has the ability as it slaps in, you know, not just the
vertical force but the rotational force as it slaps into this
panel we have adjusted the angle of the shot to try
to accommodate that, to try to add that extra force in. We
don't know enough about the range of sensitivity of this material
to be able to give you a more definitive answer?
MR. WALD: (Inaudible).
MR. HUBBARD: I'm sorry?
MR. WALD: Probably won't find that out (inaudible).
MR. HUBBARD: Yeah. What this test is trying to do is
trying to say, you know, how would you create local force?
You know, I mean, this is oh, by the way, somebody
asked what's the difference in size. You can see that this
panel let's take panel six out here, and this is panel
eight. You can see the relative difference in size, and I
went and looked it up. This is about 19 inches by 19 inches,
19 inches wide by about 19 inches tall. This one is about
28 inches wide by 25 inches tall. So, that gives you a sense
of I don't know who asked the question of the difference
in these. And what we're trying to do is determine, since
many other pieces of evidence say this is the most likely
breach location, how do you transfer significant force right
here? And so, that's why we've picked the conditions we have.
MS. BROWN: And I should point out that this model was
made without the T-seals just to show the general size and
the area of where the panels are located.
MR. HUBBARD: Right, yeah. This is not exactly representation
only. For purposes of clarity.
MS. BROWN: Richard?
MR. RICHARD HARRIS: Yeah. Richard Harris from National
Public Radio. You mentioned that panel eight, that when you
test it, it's coming off of one of the Orbiters. What about
the adjoining panels? Do they all are they essentially
going to dismantle, cannibalize from the Orbiters to do this
and, when they're done, do you would those panels go
back on the Orbiter, or will have to re-manufacture all of
MR. HUBBARD: Tested panels are not going to go back
on any flight vehicle. The panel nine is there are
about four spares in the fleet of panel nine, so that is not
an issue. We have a panel with the same kind of flight history,
that there you know, is and I think that panel
was taken off of Discovery, but, it has a spare in its place.
The decision process to use panel eight was complicated by
the fact that there was only one spare in the fleet. So, although
they interchange these panels, in effect, we are using that
spare. But, as I said before, there is another one on order.
The lead-time is something greater than six months, probably,
to produce one. But, it's our sense that that probably corresponds
reasonably well with return to flight. So, it the fleet
should end up with another spare in time.
MR. HARRIS: And 10?
MR. HUBBARD: Number 10 I believe also came off of Discovery,
but it's also one where there's multiple spares.
MS. KRISTY NABIELSKY: Kristy Nabielsky (sp) with N.K.
(sp) newspaper. If you could remind us at what angle the hot
gases hit the Orbiter, or where is it the underside,
the upper side the hot gases hit the Orbiter upon entry?
Would it be in the same place where you found the crack in
this last test, the eight-inch crack?
MR. HUBBARD: Right. We have been shooting the foam
at the bottom side, which is where the visual evidence says
that it hit. That's what you see on the film in the video
from the accident. And we created a crack in the first shot
against the RCC down here at the bottom of the wing. And it's
down in here somewhere that the breach occurred that let in
the hot gas.
MS. NABIELSKY: (Inaudible) upon entry, the angle of
the entire Orbiter?
MR. HUBBARD: Oh, it's at an angle of about 40 degrees.
MS. NABIELSKY: (Inaudible).
MR. HUBBARD: Yeah, it's coming to the underside of
the wing well, and then it (inaudible).
MS. BROWN: Okay, Gwyneth?
MS. GWYNETH SHAW: Gwyneth Shaw with the Orlando Sentinel.
You mentioned kind of a general lack of understanding of how
these cracks or damage would propagate during the rest of
the ascent during orbit and the early stages of re-entry.
Have you tested panel six to kind of see what happens to the
damage on that? And if so, how long does that take, and how
would that affect your results from panel eight?
MR. HUBBARD: The process of evaluating panel six after
it was tested, it was thoroughly evaluated prior to the test.
The panel six CAT scan and it is the same process that's
used in medical diagnostics has just been completed.
They did not find the people that did that work. Any anomalies,
any unusual features in the panel other than the crack that
was created as a result of the test. Given that that took
days to do, and we wanted to get the best pre- and post-comparison,
we've been very careful not to take the panel and flex it
or torque it or, you know, in any other way, you know, cause
it to break, because we wanted to preserve that. The evaluation
will go on of how that crack could propagate well-past the
report of the board, and I think it's part of the ongoing
understanding of how these can fail and in what way with how
MS. SHAW: Have you done anything to subject it to forces
that you would see in the rest of the assets? I mean, in terms
of how much if it's a very small crack at the time
of the strike, what happens to it after that?
MR. HUBBARD: No, we have not done any testing on that.
We've just like I said, we just completed doing the
MS. BROWN: I think one thing I should point out is
the Admiral had said that, when the report comes out, it will
be the final report, but some of the appendices may take several
weeks to follow that report. So, when the report is put out,
we don't expect to have all of the supplementary material
available on the day we release the final report. It may take
several weeks after that to release, for instance, some of
the final reports on the testing and evaluation, and some
of the other appendices. So Paul?
MR. RECER: The difference in the availability of spares
for nine and 10 versus eight suggests that eight has had a
higher rate of replacement than the other panels. Is this
true and, if so, why?
MR. HUBBARD: No. Panel nine is the one that gets replaced
the most. I mean, only three panels on Columbia, for example,
were ever replaced. They had gone in for examination and for
a certain level of maintenance and repair, but only three
panels were ever replaced. So, panel eight I can't
speak to why there's only one spare in the fleet. It may have
just been the result of focusing on the other panels that
were in, you know, higher heating area or something.
MR. RECER: (Inaudible).
MR. HUBBARD: It's the highest heating area. Nine is
the one that sees the highest temperatures.
MS. BROWN: Alan?
MR. LEVIN: I'm sorry, (inaudible). The RCC test article
is now 100 percent RCC?
MR. HUBBARD: Correct, yes.
MR. LEVIN: So that you can assess the system effect?
MR. HUBBARD: Correct, yes. I didn't make the clear.
That was one of the major findings of these last two tests
we did, was to determine that there is a system effect that
we have to account for, and that the panel eight is the target,
and that to really get it, since the models have not been
predicting the breakage very well, although predicting the
stresses pretty good we really need to have this be
as much like the real thing as possible. Therefore, it's all
RCC, T-seals, panels, everything. Five, six and seven were
not in the equation any longer. Did I answer the wrong question?
MR. LEVIN: (Inaudible).
MR. HUBBARD: Yeah. They'll well, they're going
to be downstream from the impact I mean, upstream from
the impact. The impact's going to be for the next set
of tests is going to be right here, and the force goes this-a-way.
But, they will be there, but not RCC. These will .
MR. LEVIN: (Inaudible) some anchoring from the
panels upstream (inaudible)?
MR. HUBBARD: It's on the slip side, so, yeah. Now,
they the connections that we've seen have been the
transfer of force downstream.
MR. LEVIN: With (inaudible)?
MR. HUBBARD: Upstream, correct.
MS. BROWN: Okay, Mark?
MR. CARREAU: Thank you. I'm Mark Carreau, the Houston
Chronicle. A couple of times, it's been mentioned that panel
eight is unique. I don't know if you even want to use that
word for any of them, but it's got a trapezoidal shape. It's
where the wing starts to really make that cant outward. Is
that a factor in any of your assessments that you're making?
Is there anything from the physics of the different shape
and the translation of forces just through it that are concerning
that might have made it more vulnerable, or anything in your
findings so far?
MR. HUBBARD: I don't have any really analytic answer
to the question. All I've got are the kinds of things that
the structural engineers and the physicists spat around and
talked about. We have noted the unusual shape, the fact that
this is the largest panel and, therefore, given the structure
of Columbia, as the largest distance between the supports,
it's not supported all the way across this 28 inches or so.
And that you know, if you just think about, you know,
breaking a board, you know, if you've got two supports very
far apart and you put that impact right in the center of it,
you can create a far greater deflection than if you move the
fulcrum you move the support point close together.
That's as far as our thinking has gone, although they're creating
good models mechanical engineering models of this for
UNIDENTIFIED WOMAN: (Inaudible).
MR. HUBBARD: Yeah. Columbia has a slightly different
structure sub-structure than the newer Orbiters.
MS. BROWN: I'll take just a couple more questions here.
MS. DUNN: I was wondering if you had, after each test,
when you've gone up to look at the craft, if you'd had a space
helmet on and a space suit, and were just sort of floating
there looking, how visible would this be on inspection by
an astronaut? Would it be readily available? Would you have
to really spend a lot of time there trying to see what the
MR. HUBBARD: Well, I mean, you've got a couple of hypotheticals
piled on there, so it makes it difficult to give you a detailed
answer. You know, the crack on the last fiberglass test was
very, very apparent. The crack on the RCC panel six was less
apparent, but it was there. And if you took the time to look
at it, you could see it. You know, you had to be reasonably
close, but it was visible to the naked eye.
MS. DUNN: And so, you're saying that a space walker
would be able to (inaudible)?
MR. HUBBARD: I'm I don't want to speak for the
astronaut corps here, you know. I can tell you what I saw,
but, you know, cracks that are an inch long or so are something
that you can walk up and see. It's not anything hidden under
MS. BROWN: Okay. Todd?
MR. HALVORSON: Todd Halvorson of Florida Today, and
this may be a stupid question, but it seems to me that the
like the larger cracks and damage that you have found
have been on internally, on the inside of the RCC.
Is that correct?
MR. HUBBARD: The one RCC test that we've done on panel
six, there was a three-quarters of an inch, or an inch or
so, visible on the surface. And then, when we took the panel
off and looked underneath, we found that it continued all
the way around, you know, for a total length of about 5-1/2
inches. Now, cracks that we've seen with these recent adjustments
of the clocking angle and of hitting lower down have been
seven inches long on the outside. So, I think it's too
I think you need to you know, I would stay tuned to
see what happens on panel eight.
MS. BROWN: Okay, I think that's it. Did I miss anybody?
Okay. Thank you very much. Thanks for all your patience, and
the next press briefing is July 8th, and it's going to be
earlier in the day than we usually have them. It's going to
be at 10:00 a.m. instead of 1:00 p.m.
Back to Event