Transcript – Columbia Accident Investigation Board Press Conference Tuesday, April 1, 2003 (Part 2)

So over to you, Roger Tetrault.

MR. TETRAULT: Right. I represent the technical committee; and, of course, we actually are the largest of the groups. We have five members; and those members include Sheila Widnall, who was at Langley yesterday, watching some of the wind tunnel testing that they’re doing and getting updated on that wind tunnel testing. In addition to that, we have Jim Hallock, who is at Kennedy today; and he is reviewing the debris or looking at the debris. We also have Scott Hubbard, who is working the Southwest Research testing plan; and that plan is in a little bit of flux as of today. Also we have Dr. Doug Osheroff and, amongst other things that he’s taken on, one of them is to try to come up with mechanism for chemical analysis of some of our debris to sort out at what altitudes various events happened, using Nobel gases. So that’s a very tricky kind of area that he is helping us with.

As most of you are probably aware, I am the person who is responsible for both the leading edge of the wing and also the management of the debris and its analysis down at Kennedy. Today what I’d like to do is update you with four different areas; and I’m going to update you on flight data, including the OEX, and also one small item on the old telemetry data line. Then I would also like to give you a little bit of an update on the photo analysis and what’s going on in that particular area, talk briefly about the orbiter aero off-nominal results and the struggle that we’re having to try to come to some conclusions about the meaning of those results, and then show you one piece of the debris which is, in fact, at least answering one of the questions that we have on these off-nominal aero results.

So if we put up the first slide, this one you’ve seen before; but it is the OEX recorder which, as you can see, is in pretty good shape. Let me give you a kind of verbal description of what’s on the recorder and what we can get out of it. It has two types of data. Data called PCM, which is pulse code modulation. This is low-frequency data and normally is used to get information about pressures, temperatures, and strain. It has FDM data, which is frequency division multiplex data, basically wide band. That data includes stress and strain sensors relative to the structural area, the vertical fin, the speed brake, and the heat shield, and also contains data relative to the main engine vibration.

The tape contains both ascent data and re-entry data, and I think most of you are probably aware that they turn on the OEX recorder 15 minutes prior to launch and it’s kept on until main engine cutoff plus six minutes. On reentry, it’s turned on ten minutes prior to entry interface; and the data that we have goes to the end of the tape.

Last week Scott Hubbard talked about what we could expect to find, and he told that you that there were 721 sensors that were recorded on this recorder. Of course, we’re in very much a preliminary sort of stage on the information that we’re receiving out of the recorder, but let me kind of put these 721 into buckets so that you have an idea of what kinds of data we’re seeing.

Over the weekend there were over 100 people working at NASA to try to extract this data. They were able to extract good data from approximately 420 sensors, all on the PCM side. There were approximately 50 sensors that we don’t have good data from and we don’t expect to have good data from because they were not necessarily hooked up or for a variety of other reasons they were not essentially ready to take data. There was, over the weekend, 150 sensors — and these were all strain gauge — that were on PCM No. 3 that we were not able to read over the weekend, primarily because of a synchronization problem. They were able to read that data yesterday and we now have 150 extra pieces of sensory data, but it is all strain gauge data.

That leaves you with approximately 100 sensors. Those sensors are in the FDM, and they have not been read as of now. Most of this or a good portion of it needs to be sent to Boeing and also to California before it can be read. It can’t be read here. So that gives you a pool. That accounts for 720 out of the 721 I just gave you, kind of putting them in buckets.

Let me tell you kind of in a gross-order fashion what we are finding from these sensors. We’ve taken a quick look at what they are telling us about ascent. That quick look on the strain gauges says that we are not seeing anything unique. Everything seems to be within family. I think Scott mentioned last week that he did not believe that we would see any foam strike at L plus 82 or something like that. We don’t appear to be seeing anything that matches that L plus 82.

Let me talk about the descent. The FDM data went to Greenwich Mean Time 14, zero minutes, and 19.4 seconds. That’s 15 seconds after any of the other sensory data that we previously had received. By that I mean the 32-second piece that we had that you may recall had a 5 seconds, 25 seconds missing, and then 2 final seconds. So we’ve got 15 seconds more than that now. In addition, we have sensors that now tell us about the 25 seconds which previously was missing.

The PCM data did not quite go so long. It went to Greenwich Mean Time 14, zero minutes, 13.4 seconds, or 9 seconds longer than the previous one. What all this suggests, of course, is that the OEX recorder, which is inside the fuselage, was receiving power during that entire period of time; and, of course, at that point the fuselage had probably not broken up.

Let me put another slide up, if I could. This is a very preliminary slide. It’s under revision, but what I wanted to show you is I wanted to talk about four sensors and give you some information about four sensors. One is going to be on the OMS pod, which is not shown here. This is the lower portion of the left wing. The three that I want to talk about are all going to be in this area. One sensor is not shown. It is a strain sensor, and these are two temperature sensors.

This is 9910 and 9895. 9910 is in front of the spar. 9895 is behind the spar. Now, let me talk about these sensors and how they affect the beginning of the time line for events that we’ve seen before.

The first one I want to talk about is not shown on here. It is Sensor G9921, and it is right in this area behind the spar, at approximately RCC Panel No. 9. As I said, it is a strain gauge. It starts to increase and goes off nominal at Greenwich Mean Time 13:48:39. That’s EI plus 270 seconds; and I’m going to try to give you, where I can, both the Greenwich Mean Time and the entry interface plus seconds after entry interface, because that’s how these gauges are measured.

As I said, it’s located behind RCC Panel No. 9; and it did go off nominal at 270 seconds. That’s 206 seconds earlier than the first event on our previous time line. And the first event, you may recall, on the previous time line was the start of the delta yaw which was at EI plus 476 seconds.

The second one that I want to talk about is Sensor 9910, which is the one in front of the spar. It is actually attached to the clevis. I believe it is the clevis which is the outboard attachment for the RCC Panel No. 9 and where it mates with T Seal No. 10. So it’s on the outboard side of RCC No. 9, forward of the spar. That one went off nominal at 13:48:59. That’s at EI plus 290 seconds. It went off signal low at EI 492 seconds or approximately 52 minutes and 19 seconds. When it went off signal low, it had read only about 50 degrees, which probably suggested that the wire had been cut; and that is not particularly surprising when you recognize that this sensor actually has insulation which is packed around it, which are referred to as earmuffs.

Let me talk about the third sensor, and I’m giving you these in order. These are the four sensors in order as they went off on this OEX recorder. The third sensor is Temperature Sensor 9220. It is located on the OMS pod TPS surface and on the leading side of the OMS pod. It starts to go off temperature and it goes low off temperature at 13:49:53 Greenwich Mean Time, which is entry interface plus 344 seconds. It goes off low until approximately 540 seconds, and then it makes a rapid switch and goes high. And it actually goes as high as 1200 degrees.

Normal for that portion of the flight would expect to be around 600. So what can you gather from that? First we think that there is a change in the mass flow which moves the mass flow probably below the OMS pod and then there is something else which happens that raised the temperature rapidly up, which may be burning or burning aluminum or a number of other effects that we have to go look at.

The fourth sensor is 9895. This is the one behind the spar at RCC Panel No. 9. This one right here. And it’s pretty much in the middle of the wing surface in terms of where it’s positioned. It goes off nominal at 13:51:14, or EI plus 425 seconds. It goes off scale low at EI plus 520. So the wire is cut at approximately 520 seconds.

This particular one behind the spar goes up to approximately 450 degrees where it peaks and then goes off scale low. This is 63 seconds before the old time line Temperature D sensor, which was the first sensor in the old time line that was a temperature sensor. And if I go back and look at Event No. 2 which I gave you, which was the one in front of it, that was 188 seconds before brake line Temperature Sensor D first showed an off-nominal.

Now, what I would like to mention to you is it’s probably even more of a delta than that because we are finding that brake line Temperature D, which established the first sensor previously, is probably not off nominal at the time that’s on the time line. It’s probably later than that. That’s an important point strictly because Sensor D goes off first and it is in the aft of the wheel well and it is high in the wheel well, which tends to indicate that the heat is coming from the aft part of the wheel well, moving forward. But if you really look at it and say what’s going on and that moves, it is much more likely that the heat is coming from the forward side of the wheel well going aft and this study that they’ve done that indicates Temperature Sensor D is probably nominal for a much longer period of time than they specified helps in making that case that this all begins to line up a little bit better than it had previously. In that study which NASA ran, they found that at least 13 flights had had bit flips before EI plus 500, and 11 of them had been before this flight showed the same kind of change.

Okay. Let me kind of give a general summary of the data that we’re seeing. This is very general and very early, just as giving you some information about some of these individual sensors are, without a lot of review. In general, there are about 15 sensors in the left wing; and they all seem to go off scale low about Greenwich Mean Time 13:22:54, which is west of California. NASA has also looked at 17 sensors, all on the left wing, and has indicated that they believe they see a warming trend compared to previous flights with similar inclinations, and that warming trend — not off-nominal trend, just a general warming trend — occurs at EI plus 80. Very early in this flight.

Let me have the next slide, please. Scott showed you this slide last week, and what I wanted to do is update you on some photo analysis which has been done. NIMA, which is the National Imaging and Mapping Agency, has taken all of the photos that we had and digitized them at 50 megabytes per frame. Having done that, we have been able to actually come to some conclusions, and let me say what those conclusions are.

ADM. GEHMAN: Roger.

MR. TETRAULT: Yes, sir.

ADM. GEHMAN: We have to hurry up.

MR. TETRAULT: Okay.

ADM. GEHMAN: Fifteen minutes has already gone by.

MR. TETRAULT: I apologize.

REPORTERS: We’re happy.

MR. TETRAULT: The important issue here is the debris size, which has now been calculated to be 24, plus or minus 3; 15, plus or minus 3, and 5, plus or minus 1 — which equates to approximately a 2-pound piece. The 24 and the 15 are measured; the 5 is a calculated value based upon the trajectory that it flew and the mass that it would have had to fly that trajectory and wind up with a velocity of 640 feet per second at the end, which is the average of the measured values that they have that it hit the wing.

There is no evidence of more than one strike. The impact point as you see it here is between RCC Panels No. 5 and 7; and no debris is seen over the top of the wing. And that photo that we previously showed you with before and after debris strike, with 16 frames and 17 frames and showing no damage to the acreage tile is now done at 30 frames and we still see no damage to the acreage tile. So this portion of the likely area that hit is a little bit suspect.

Okay. Let me just go quickly here. Next slide. This is the anomalous aero moment that we’re looking at and are struggling with. The one that I like to call your attention to — and this chart, by the way, was put together by Sheila Widnall to try to get the trajectory matching the aero that we’ve seen — this represents loss of lift in the left wing and then additional lift in the left wing, which we have not been able to explain. So trying to explain what’s going on here has been difficult.

If you were at the public hearings, NASA showed some wind tunnel tests that indicated, if the door was open and the ear was down, that it could happen. Other things that had been postulated are aluminum burning, missing bottoms to the RCC, a jet flow out of the wheel well, and some kind of flow on the tail.

Let me go to the next slide. I just want to show you that, given the debris that we’re looking at, one of these, which is that the left main gear had deployed early and is creating an aero moment, is not likely to be a correct theory. This is a wheel, the wheel strut that you can see in this picture; and if you look, you can see nice, shiny chromium plating that goes on that wheel strut. If the plating was eroded in actual ballistic flight after breakup, the erosion would be the same on both sides, which it isn’t. What you can see is there’s more chromium on this side, which is the forward surface when it’s deployed. If that had been deployed, that chrome would not be there. That’s the bottom line.

So it says basically that the wheel well, as other data has begun to indicate to us, the wheel well was closed, the door was there, the wheels were up, and that’s not the cause of that moment that we’re seeing. We’re still struggling with how to get there.

That’s all I have.

ADM. GEHMAN: Thank you very much. I hope you can see from the three briefers — for example, John Barry talked about the radar reflectivity studies at Wright Patterson Air Force Base — the only piece of debris left on the table is a carrier panel. Roger’s little pork chop diagram there covers two carrier panels. You can see how much work it takes and how these things need to line up, and you can see how much evidence it takes to convince us of anything. We are very skeptical.

I apologize for interrupting Mr. Tetrault there, but he could go on all afternoon. So, with that, over to you.

LT. COL. WOODYARD: We’ll turn it over for questions. We’ll continue our same format and we’ll start here in Houston and then we’ll go to our phone bridge.

Please identify your name and your organization, as well.

A REPORTER: USA Today. So are you convinced now that it was a carrier panel that came off on Flight Day 2, because you said there was some more testing to be done and I wasn’t sure if that was only to narrow down the number of tiles that came off the panel or if you had other objects to rule out?

GEN. BARRY: We’ve exhausted all the testing on the parts that were given. There were 22 external and internal shuttle components tested at Wright Patterson, and all of them were eliminated except for the carrier panel as the leading candidate. As I mentioned earlier, there might have been some of that with the blanket, but when you compare the results of the area, the mass area ratio and the burn as it re-entered, it didn’t make any sense. So right now the carrier panel is the leading indicator of all of the 29 candidate shuttle components that we provided at Wright Patterson and the Air Force research lab.

ADM. GEHMAN: But that does not equate to we’re being convinced that it is the carrier panel. As a matter of fact, we were meeting earlier today; and we came up with a couple of other things we would like to have tested.

MR. TETRAULT: In addition, the testing that’s going to go on at Southwest Research should help us really narrow down what items are weak in this entire system and where a breach may have occurred and help either deny or confirm this piece of data.

A REPORTER: Houston Chronicle. For General Barry. Can you just go over again the major components in this area that were tested? I think Admiral Gehman just mentioned there’s a couple of more things you would like to test. Can you tell us what those are, please?

GEN. BARRY: The bottom line is that we tested, including the different kinds of tiles, the fibrous materials, the beta cloth, the insulation blankets, the carrier panel that we talked about with the horse collar, the RCC edge and earmuff seal, different variants of those, all totaling to about 22 components. We had 29 to select from, but those were the ones that we looked at for that part of the test. The eliminated majority of the components is either unknown objects either due to low or too-low RCS signatures, radar cross-section signatures. The possible candidates still stood out, and we want to look at the carrier panel now with not only three tiles but also four tiles.

If you take the carrier panel and you include the horse collar, it’s a very strong candidate, matches up with the radar cross-section. If you take the carrier panel without the horse collar but increase it to four tiles, that also gives you a pretty good indication, we surmise. We’ve got that one added test to go through. In any case, the others have been pretty much eliminated for the reasons that I have stated.

A REPORTER: New York Times. Is the foam that’s being shot at tile being tested with and without ice, or are we just looking at foam?

MR. TETRAULT: The testing right now is probably going to be delayed about a week. As I said, it was in flux; and the reason for that is this new photo information which tends to indicate a piece that’s larger than what we previously had been looking at. So now the debate is what size do we actually shoot. As I mentioned to you, the thickness dimension was 5 inches. That was calculated on the mass that it would take to fly the trajectory that it would take. So if there was ice in there — this was calculated on based on this being strictly foam. If there was ice in there which would have added to the mass, then you would have to subtract from it the 5-inch dimension to be equivalent. So to the extent that there is weight, it takes that into account already by having that 5-inch dimension.

A REPORTER: Associated Press. For General Barry. I wanted to talk a little bit about the zinc oxide contamination. Was NASA aware of this? Were they taking any steps to work around this? And just how damaging, in your opinion, is an RCC tile that got 20 to 40 pinholes?

GEN. BARRY: As you know, we’ve been examining the issue of ageing and looking at candidates for pinholes to explain them. Zinc oxide, that can come from primer as the infrastructure ages and doesn’t get painted again, with the added rain, particularly when the orbiter is on the pad and ready for launch. This seems to be the leading candidate right now. There are other explanations, something to do with oxidation and so forth. We wanted to just kind of focus on this one. This does seem to be, from the experts’ standpoint, the leading candidate. So we’ve taken that. We are doing some more study. We want to examine all of the issues in the pinholes as close as we possibly can to try to get to the bottom of the question — the question being, if something hit the RCC and it was an aged RCC with pinholes or oxidation underneath that, would that be a contributing aspect to this mishap. If it hit a brand-new RCC, could it withstand it? That’s why this testing they’re going to do with the foam will be instructive, because we’re not only going to send it against tile but we’re going to send it against RCC. Roger, can talk a little more about that.

A REPORTER: I was just wondering. Were you aware of NASA having concern about this ageing?

GEN. BARRY: Yes. They had looked at this as a possible contender, and this is one of those things they have been working on for the pinhole explanations. Remember, pinholes that are less than .04 inches were greeted as okay. And they just went about either refurbishment and it was out of tolerance or they went on a repair, as Steve Turcotte mentioned last week — as he’s our leading guy on the RCC, he and his team. So they’ve gone through and listed all of the different RCCs that have been either refurbished or repaired, and I can give you that data again if you want it.

A REPORTER: ABC News. Mr. Wallace, you mentioned that you think we’re not really quite keeping the E-mails in perspective. Would you care to put them in perspective for us, sir?

MR. WALLACE: I think I would just say again that they are part of a very complicated story and you haven’t seen the rest of the story. A lot of it is developed in the context of interviews where they are strictly privileged; and this being a safety investigation where our sole objective is to determine the probable cause and look for ways to raise the level of safety, we grant these witnesses privilege, which means that their identity and what they say is protected, yet we use that information to form a part of the entire story.

Of course, we have done extensive review of material, much of which will be available to you as well, things like records of meetings from NASA and actually listening to tape recordings of meetings and comparing them with the written records and trying to sense the whole dynamic. So sort of a brush stroke by brush stroke, it gives you a picture which we have not completed; but I would just say that you don’t get a balanced picture or a complete picture by any means by seeing the E-mails.

A REPORTER: Orlando Sentinel for Mr. Tetrault. Can you give us a sense of where the analysis of the OEX recorder data will go from here on, what kinds of questions and what kinds of data will be addressed?

MR. TETRAULT: Well, the kinds of data is just going to be temperatures, pressures, strains, some vibration, those sorts of things. What it is obviously showing, though, is events that are occurring much earlier in the time line, although my general sense is that it’s not showing us anything that we didn’t suspect that it was going to show us. It will show us, I think the one thing that I see, is that when we start looking at where the lines were cut, the electrical lines were cut, part of them were cut around 540 seconds after EI, that it will begin to help localize where in the front edge the problem is. Up until now, if I took collectively all the data that we had, I could in my mind locate this event somewhere between RCC Panel No. 5 and No. 12. I think this will help shrink that down to some much narrower one. Then when you begin to put that where the photos are showing where the foam hit, then I think you can start making some assessments of what are the probabilities of that being an initiating event, if you will. I wouldn’t say necessarily the cause, but there may be an initiating event that leads to a whole series of other things that lines up with the accident.

ADM. GEHMAN: Roger, if I don’t have this exactly right, I believe that, in answering this question, that this data is being looked at by literally hundreds of engineers from different disciplines and I believe that what will happen is that we will put out revisions, more revisions to the time line. We’re up to Revision 15 already. I think that periodically as this data is analyzed to a degree with the engineers confident what they’re looking at, we will continue to put out Revision 16, then Revision 17 and Revision 18. As you may be aware, when we put those revisions out, we annotate what’s new on it. That’s my understanding.

MR. TETRAULT: That’s exactly correct. In fact, there is a preliminary time line that has some of these sensors already on it, but that’s what they’re working on and some of this data they only received yesterday and some of it they still don’t have. But in looking at it, there’s a lot of information on strain and stress that isn’t going to show you a lot.