§1. Cardiac guide wires and cow magnets [00:04]
We're talking about stainless steels, and this is a guide wire. Has anybody ever had heart surgery? They come in through a big artery in your leg and snake these little stainless steel wires, teflon-coated, up through the vein. This is an old one, about twenty-five years old. The thing is very flexible at the tip. The surgeon takes it through the arteries, gets it where they need to go, and then they use it as sort of a transport rope to send other things up there — a balloon, a stent, whatever they want to put in your heart.
They were having problems with the weld breaking. It only had to lift about two and a half pounds, and it wouldn't even do that. This is nominally austenitic stainless steel 304, which we all think of as fairly non-magnetic. It's a very fine wire coiled to make it flexible — like a little coil spring with another wire down the center.
I actually got two new ones because there was a lawsuit. Some guy supposedly got an errant operation and they left about thirty inches of it inside him. They sue anyway. These cost $1,500 apiece.
[Tom produces a magnet.] It's a stainless steel magnet. None of you grew up on a farm, I take it. This is a cow magnet. You make the dairy cattle swallow this. It turns out, when they're out grazing, they sometimes pick up wire from the fence. If that gets into their udders, it destroys them, and they're not much good as a milk cow at that point. So the stainless steel magnet grabs onto the carbon steel that they might eat and holds it in the stomach. The stainless steel doesn't corrode in stomach acid, but the carbon steel sticks to the magnet and stays put.
That's why 7-Up was named 7-Up, by the way. It had lithium in it — you could buy it as a soft drink in the 1930s. Lithium is an antidepressant. And Coca-Cola — Coke was invented by a druggist in Georgia at a pharmacy soda fountain. The Coca-Cola formula is one of the most tightly kept secrets in the world, in a vault in Atlanta.
§2. Stress relief in welding [04:45]
I want to mention some things about stress relief. This applies to all welds, not just stainless. You don't want to stress relieve stainless steel because you'll be right in the sensitization range — you can destroy the corrosion resistance with a thermal stress relief. But stress relief in welding is needed. If I have a joint with no included angle — an electron beam weld, say — there's almost no distortion; the metal shrinks a little because it contracts on solidification. But as I open up the angle, I get a V-shaped weld, and the bigger the V, the more contraction. Steel is about three and a half percent contraction on solidification; aluminum is six percent. Aluminum has much bigger distortion problems than steel.
You have to balance welds. If you weld a groove weld all from one side, then go fill up the other side, it never comes all the way back. You're better off welding partway on one side, flipping, welding partway on the other, and alternating. One of my students, John Galati, his first job at Electric Boat was developing ways to weld the torpedo tubes — about forty feet long. They had to weld several tubes together, and he balanced the welds on either side as he was going around the circle. You might do a weld at three o'clock, change, go to nine o'clock, balance the welds around the circumference to get rid of the distortion.
You have a choice in welding between distortion and residual stresses: you can have residual stresses and no distortion, or a lot of distortion and no residual stresses. It's more complex than that. Thicker materials are so rigid that once you start building up thickness — say you're welding something four inches thick — once you put a two-inch weld in there, it's not going to bend anymore. It's a solid rigid structure. Whereas thin material — sheet metal, eighth-inch material — you weld one pass and it distorts. Thin material tends to distort, thick material tends to have residual stresses. The worst thickness of all, the most difficult to deal with, is about three-eighths of an inch or one centimeter. You surface ship guys — that's the thickness you tend to be dealing with.
You get a fair amount of distortion and you can also get residual stresses. And now people are making eighth-inch steel ships. Here are actual pictures of welds. This is 5083 aluminum with a fairly deep, parallel-sided penetrating weld — no distortion. When you get parallel-sided welds, you can avoid distortion. As you get thicker and the weld opens up, you get more and more distortion.
§3. Post-weld heat treatment and mechanical stress relief [08:57]
When you have residual stresses, you lose fatigue strength, fracture toughness — lots of things. The American Society of Mechanical Engineers, if you're building pressure vessels, requires you stress relieve anything above an inch and a half thick. The hull of a sub is more than an inch and a half thick. How do we usually stress relieve welds?
Student: Post-weld heat treatment.
Post-weld heat treatment, for carbon steels and alloy steels. We pre-heat steels to drive the hydrogen off. We post-heat them to soften the material if it's a highly alloyed steel. But once you get to inch and a quarter, inch and a half, you're going to have yield-level residual stresses in there. The material has plastically deformed as you continue to put weld bead in. Go four inches thick, I guarantee you have residual stresses equal to the yield strength of the material. In a complex three-dimensional structure you can have triaxial stresses. You can have HY-100 [steel] with stresses of 120, 130 ksi in a complex shape.
So typically for things inch-and-a-half thick, ASME boiler pressure code says you must stress relieve to meet the code. Submarines especially. Note: that's post-weld heat treatment in the eleven-, twelve-hundred-degree range, not the three- to four-hundred-degree preheat to keep hydrogen out. You can preheat and you can low-temperature post-weld heat treat with little blankets right around the weld to diffuse the hydrogen out. You don't stress relieve the whole vessel.
For thermal stress relief, this is a pressure vessel shop. Here's a vessel, and they're putting blankets all the way around. After they weld the whole thing, they heat it up to 1100 degrees Fahrenheit, typically per hour of thickness. They build a furnace around it. The blankets are like welding power supplies. Surface ships, you tend to use electrical heaters. I don't know what they're doing in the service shipyards right now. But you build it, you put insulation, you build your furnace around the vessel.
In some places they have huge vessels — Babcock & Wilcox or somewhere — with a heat treating furnace the size of a two-story house, and they put the whole vessel right inside that furnace. This other one is portable. A company comes in and builds a box about the size of a railroad car, on railroad tracks. The furnace slides over the vessel. You heat it up to 1400 degrees, might take a day or two to heat the whole thing, and hold it there for a couple of hours.
How do you stress relieve a sub? You do it on the first deep dive. It's mechanical stress relief. It's full of residual stresses until you go down deep, and the whole thing gets squeezed. The areas that are in high tensile stress, when you put them in big compression, they yield. You have to have equilibrium of forces — for all you mechanical engineers, you have to have equilibrium of forces. If you have tensile residual stresses here, you have to balance them with an equal amount of compressive forces. So if you compress at a fairly high stress level, the compressed areas yield. As they yield, when they come back up, there's less residual stress. You can mechanically relieve residual stresses.
I'll talk about this more when we get to aluminum alloys. The Davenport, Iowa plant of Alcoa where they roll heavy plate — they stress relieve mechanically. Six-inch-thick plate, ten feet wide. They have huge hydraulic jacks in a room about half the size of a football field. They grab the plate and pull it three percent. Figure out the force: 70 ksi yield, ten feet wide, five or six inches thick — millions and millions of pounds. A lot of force to pull these things apart. But getting rid of residual stresses is something you need to do.
You can also peen. We don't build battleships anymore, but they used to peen them. If I went back to my old Burstal and Adams Welding Metallurgy book — Burstal was the guy who wrote about weldability of steels in the 1940s — that little list of things, it says "peening necessary." You fill up the bottom part of your weld and you come in and beat the surface to relieve the residual stresses.
§4. Peening titanium and the CF-30 in the Vietnam War [16:15]
When I worked at the Naval Air Rework Facility one summer between my freshman and sophomore years, they had a CF-30 engine — middle of the Vietnam War, they needed to get these back out so they could get shot down again. They had a shortage of CF-30s. They completely rebuilt the engine, it was in the shipping area ready to go, and the final inspector says, there's a crack on one of the vanes on the end of the compressor, made out of titanium. Filled up with a black goo plastic for dampening purposes.
They decided they didn't want to take two or three weeks to disassemble and rebuild it, they needed it out in the fleet. So the boss — Roy was his name, he had twenty engineers, and I was just a summer intern, the low man on the totem pole — Roy says, you're going to handle this weld repair. He explained we were going to use an Almen gage. Almen — a General Motors guy, look it up, developed in the fifties or sixties. You take a little strip of steel, a sixteenth of an inch, and if you peen the surface — like shot blasting — you get a curvature from the residual stresses. The intensity of the shot peen is proportional to the curvature.
So we made up a little tool with a hammer on the end of a vibrating air hammer, found this old ball-peen anvil, and he would cover the surface, peening it. I was supposed to sit there and measure the height of the Almen gage. We were setting up a procedure to do the titanium. The data was all over the map. Press soft, press hard, do it for fifteen seconds, do it for thirty seconds — the heights were all over the map.
Then: how do I do it on the titanium? They had a guy weld it. You can see the weld bead. We can't grind it off, we're just going to leave it. But we had to do the peening. Now, scientifically, on a flat surface it's a bulge — he's got a weld bead reinforcement on what was thin material, just a thirty-second of an inch reinforcement, but now he has to peen an uneven surface. That's going to make it more inconsistent.
So I had to go out, representing the engineering department, and watch this guy peen it for a certain amount of time. Then I came back in and Roy says, sign this. I said, what's that? He had a signature line for two signatures, and he says, that's signing that we've repaired this weld. I didn't know anything about welding at that point. I said, what happens if I sign? He says, if this plane goes down, it'll be a deal within twenty-four hours.
That's how things really get fixed when you need to fix them. Roy had been in the business for forty years. We were taking some risk, but it wasn't a big risk. I've learned since and realized it wasn't really that big a risk, but I didn't know that then. I trusted Roy. If I had to do it today, knowing what I know about fracture mechanics — we just rewrote our liability laws for professional engineers. We have a statute of repose now: engineers can't be held legally liable past, I think, twenty-five years.
§5. Professional liability and the loose bolt [21:34]
In aviation, places like Cessna would have all gone broke because of lawsuits any time a pilot did something stupid. The Federal Aviation Administration has done lots of studies on why private aircraft go down — eighty-five percent of the time it's pilot error. Most of the rest of the time it's the A&P, the airframe and powerplant mechanic. Mechanics make mistakes too.
Just last week or so, someone called me up. I'll write a paper, I won't say too much, but it's a piece of farm equipment, about $700,000, only had a hundred hours on it. A bolt came loose, and everybody agrees that when the bolt came loose in the engine, it started a fire and destroyed the $700,000 piece of machinery. They moved it forty miles on a flatbed truck to a shop to do a teardown inspection. Everybody agrees, this bolt allowed the diesel fuel from the engine to get out, found an ignition source, and started a fire. The guy was riding the machine through the fields, got out, tried to put it out by throwing dirt on top of the engine.
The attorney for the defense of the manufacturer's theory is, when they put it on the flatbed trailer to move it down the highway, that's when the bolt vibrated loose. I said, that's a really good theory — they've now changed an assembly defect into a design defect. They're saying, this vehicle is designed to go over the fields with the engine running and vibrate, but going down a highway with the engine not running, the bolts come loose. So you have attorneys who will come up with anything.
We do have, in aviation, the General Aviation Revitalization Act — GARA — from the early nineties. GARA says any part that's eighteen years old or older, you cannot sue for a defect in manufacture. A lot of states have laws that if you buy a manufactured product — a washing machine that lasts ten years — you can't sue for a manufacturing defect. You can still go after a design defect. We do have laws that limit liability, but there's not one that protects professional people that I know of.
One time someone wanted me to work on a commercial job, no litigation, but I had to have professional liability insurance. I went to look — about $500 worth of liability insurance, good for one year. If I ever got sued, I'd have to maintain that insurance for the next fifty years. So to do a $20,000 job, I'd have to pay $50,000 over fifty years for liability insurance. I decided not to get insurance.
One thing is, I never design a part. I can tell what's wrong, and I give some options for what designs could be, but I don't make the final choice. One time I developed a surgical instrument consulting for Johnson & Johnson. They do laser surgery, and metal instruments can bounce the laser off and burn the patient somewhere they didn't want to burn. I came up with a surface coating that absorbed the laser light without reflecting, and we got a patent on it. If Johnson & Johnson ever got sued, I'm not an employee — I could be sued personally. I ended up writing to Johnson & Johnson asking to be held harmless. It took six months in their legal department. They finally gave me a letter saying we'll treat you as if you were an employee — my legal fees would be paid.
You've got to be careful about professional liability. I have an LLC, a limited liability corporation. You hire the LLC, not Tom. I don't keep a lot of money in that account; I transfer it as income to personal. The other thing — I own a condo. If you register your home as a homestead at Middlesex Courthouse, sign a little form and register it with the clerk of court, if you ever get sued for anything — some neighbor's kid falls on the trampoline — they can take all your assets except your house. The government doesn't want to make you homeless. So there are some laws that protect people, but there's powerful lobbying in Congress not to reform a lot of the laws. It's a huge business — tens of billions of dollars in lawyers' fees.
§6. Fatigue design and the Aerostar axle [30:00]
Let me go over fatigue design. Designing for welds that might fatigue — you have some structure that's moving, the welds could fatigue. This is primarily for steels, but it also applies to other alloys. There's a book written on fatigue of welds. Mostly steels. If you look at just a solid bar of steel, the fatigue stress range goes up with the strength of the steel. At 900 megapascals, fatigue is about a forty-five-degree slope — the fatigue strength of a simple solid bar of steel is proportional to the tensile strength.
If you put a hole in it, it goes down by about thirty percent. If you put a weld in it, no matter what the strength of that steel, it's going to behave like it's a low-strength steel. Putting welds in a steel structure will bring everyone back to the equivalent of a mild steel.
To give you an example, in 1985 Ford came up with the Aerostar van, and they were using high-strength low-alloy steel for the first time instead of plain carbon sheet steel. They designed the rear axle to have a fatigue strength proportional to the tensile strength of the better steel — that's what their designers knew. That would be true if it weren't welded, but the rear axle had brackets welded onto it, and they found people throwing axles, fatigue-failing them in the first six months. This was not a great day for Ford. Big product recall. There was a tremendous amount of work done after that on the fatigue strength of sheet metal welds in high-strength low-alloy steels.
You always have a little stress concentration at the weld. Statistically, this area may not be bad, this area is worse — there's always some bad area. You've actually got several types of stress concentration. Geometric stress concentration at the notch — say it's a fillet weld, you've got geometric stress concentration at both corners. Metallurgical stress concentration because you have a different microstructure in the heat-affected zone. You can also have slight undercutting, or as the weld solidifies you can have microcracks, hot shrinkage — less than a human hair, but enough. When all these things line up at the same spot, it doesn't do you any good to weld.
§7. Truck frames and cement-truck water tanks [33:58]
If you look at the frame of a tractor-trailer going down the road, that's a big plate of steel made into a C-channel. The steel frame is ladder-frame construction — a C-channel here, another over here, those are the two ladder rails. Usually a high-strength steel, eighty or ninety ksi. You are not allowed to weld on that. National Highway Transportation Safety Authority — you're not allowed to weld. Some old boy comes along, wants to add a little bracket to the bottom of his truck. He puts a little weld down there, and 100,000 miles later — which is not a lot on a truck — the frame falls apart. You can't weld high-strength steels and keep the strength. You can bolt on it, they put drilled holes in for that, but you don't weld on it.
Automotive is unibody construction. Some pickup trucks don't have the stresses of a big eighteen-wheeler and use welded ladder-frame construction. The Dodge Ram, Chevy Silverado — some of those are welded. But they're welded under very carefully controlled conditions in plant, and they've done fatigue tests so the welds are not in the most stressed locations. On the big eighteen-wheeler frame trucks, I've never seen anything except bolted connections.
You have lots of different geometries of welds. This is one I'll show you later — a cover plate on something, and you're not allowed to weld around the corner. You weld just the two sides, leave the corner alone, because of residual stresses. There are sometimes hard spots — on board ship you might have something holding something else and the weld stops right here. That's a hard spot, a stress concentration, and that's where the failure starts. You have a lot of these on ships.
On pressure vessels you put a doubler plate on the shell, and the doubler plate is welded radially. Much lower stress because you're spreading the load over a larger area, not concentrating it on one spot. So there are lots of details you can use to reduce the stress.
I had a situation once on cement trucks. They have a 200-gallon tank of water, and they use air pressure from the air brakes — the truck has 55 psi air to run the brakes. In some residential parts of the country it says "air brakes not allowed," because if you've ever heard a truck with air brakes, it's pretty noisy. So you have this 55 psi air, and cement trucks always have people spilling cement, so they carry this 200-gallon tank of water with hoses they can pressurize. About the pressure of your garden hose. They take it, spray off the truck, so they don't get cement caked on it. When they really have a problem they often carry a couple gallons of muriatic acid — HCl, hydrochloric acid — but you'll see corrosion if they don't rinse with water.
A guy came back from Iraq, was working as a welder in Pennsylvania. They had a tank with a leak. The tank was made out of aluminum. Officially you were supposed to send them back to the company, because they wanted to sell you a new tank. These things were always fatiguing — vibrating going down the road with 200 gallons of water in them, it's a big heavy thing. The ASME boiler and pressure vessel code requires doubler plates on end caps — if you have a nozzle penetrating the wall of your tank, you have a doubler plate, you fillet weld here, fillet weld here, so you don't have a sharp stress concentration. Required under the ASME code. But ASME code excludes water tanks.
The company out in Iowa making these was sort of a mom-and-pop shop. They asked the Department of Transportation, which regulates trucks on the highway, what design rules they had to follow. DOT said, that's just an appendage, that's just something you're carrying — nothing to do with the truck going down the highway. It's like throwing a water tank in the back of your pickup truck; we don't regulate what you put in the back of your pickup. Then they called Pennsylvania, said we've got this vessel carrying water. Pennsylvania said water vessels are excluded from the boiler and pressure vessel code. So no doubler plates.
The guy making this decision wasn't an engineer. The company didn't have an engineer. They were buying hydraulics from one company, tires from another, assembling components. The guy making the decision was their attorney, looking at the regulations. He had no engineering training. He made the decision: simple design, no doubler plates.
So this guy in Pennsylvania is fixing some of the welds — fractures all the time. The attorney for the company actually had a nice business selling replacement tanks. He's welding on this and goes to check if he fixed the leak. He's supposed to use 5 psi, but he had 100 psi on it — he blows the tank up. The end cap comes off, blows him fifty yards away against a concrete wall, in four pieces. He survived three years in Iraq getting shot at and didn't survive welding in Pennsylvania.
They had actually had an accident once before — the company would do a pressure test on these tanks before sending them out. They'd cut a guy's leg off when one let go, but he wasn't killed. The attorney decided it must be a fluke and ordered that all future pressure tests be done inside a cage, in case they had another fluke.
When you fill up the tire on a big eighteen-wheeler, by OSHA you must do that inside a cage — one-inch heavy-wall steel bars — because that tire goes like a bomb, people get killed standing next to it. So it turns out it was a mom-and-pop design, they didn't have a clue. The attorney didn't have a clue. I only got involved because I was working with a guy in California, Roger, who used to be on the main committee of the boiler and pressure vessel code. He called Pennsylvania. Under Pennsylvania law, this tank should have been required to meet the ASME code, even though ASME code excluded the particular application. It was a quirk of Pennsylvania law. So it wasn't a legal tank.
The real problem Roger realized was, how many of these are out there? A hundred thousand cement trucks all over the country. Roger goes to his friend, head of the National Board of Pressure Vessel Inspectors, and says, we've got a hundred thousand bombs out there, we need to get them inspected. And the guy says, we can't do it — we don't have enough pressure vessel inspectors in all the states to take on an extra hundred thousand vessels. Some of the complications when you have lawyers doing the design.
Student: Why did it fail — was the weld weakening it?
The weld wasn't great. But the real reason was, the thing was designed for 55 psi, and if you put 100 psi in it — that's the test pressure he was supposed to have a regulator on. It would limit to 5 psi, and the tank would have taken 5 psi, and they would have found there were new leaks. But he put 100 psi on it. No one's really allowed to know exactly, but we think he connected the air pressure, went off to do something else, and was standing in the wrong spot when it let go. The welds weren't great either. Repair welds were not so good. Steel isn't that hard to weld — you can learn to weld steel reasonably in a couple of weeks, in fact the shipyard welding programs run two weeks. There are some businesses where you go to school for one day and learn. To do a good job, you need six months. Titanium is harder than aluminum. It's actually easier to make a good-looking titanium weld; it's hard to get a non-contaminated one. We'll talk about that.
§8. Fatigue details and the King Street Bridge [48:47]
You've seen this detail — rat holes or mouse holes. You don't like intersecting welds. Sometimes you have to have intersecting welds, but you'd rather have a half-circle hole than two intersecting welds, because intersecting welds have tremendous residual stresses. This is out of the civil engineers' steel construction manual — there's a chapter on fatigue. You can look at loading conditions: your bridge or your building is going to see less than 100,000 cycles, less than half a million, less than two million, or over two million cycles. That's your loading condition. Then you come over to another table that goes on for pages. Plain material is stress category A. Put a weld in it and all of a sudden it's down at stress category B.
This should eventually get out — it's in the structural welding code, AWS D1.1, the red, white, and black book. They have similar things. Stress categories down to E and F, fatigue curves for the different categories — S-N curves. You can draw a picture for each. Stress category A is a built-up member, simple double butt weld in axial tension, something in bending, fillet weld built-up member; here's that doubler plate where you don't weld across the front, you don't go around the corners — I'll show you why. You can usually find a joint detail similar to something in your structure, and it will tell you a factor for the fatigue strength. For carbon steel, the fatigue strength goes from about twenty-one ksi without a weld down to maybe seven for a complex weld geometry. A factor of three reduction in the stress you can tolerate, depending on the geometry of the structure and where the weld is located.
You have to pay attention to design for fatigue. This is the King Street Bridge in Melbourne, Australia. I think it failed in the late 1970s. Cars going across the bridge, a major thoroughfare. Brittle fracture going out through here. Here's your weld detail — doubler plate, on the bottom flange, probably a cover plate. It had a hydrogen crack from welding, grew by fatigue to a critical size, and then bang. They lost a major thoroughfare for about a year. Since then you're not allowed to put the weld going across that direction — you weld in this direction and this direction, because the corners are tremendous sources of hydrogen cracks and fatigue. Certain types of details are forbidden, not because someone calculated it one day, but because something failed. That's how codes and standards are developed — it's experience. Someone has a failure, we go study why.
§9. *To Engineer is Human* and the Thresher [53:29]
Henry Petroski, a guy at Duke University in the civil engineering department, wrote a book called To Engineer is Human. He's gotten famous off it, elected a member of the National Academy of Engineering. The book is about how we learn to engineer new structures from failures. We learn from the Sea Wolf problem — expensive lesson to learn, but we learned a lot about welding high-strength steels we didn't know about. We learn about process control from the Thresher — the whole SUBSAFE program came out of the Thresher disaster.
The Thresher was built up here in Portsmouth Naval Shipyard. I was living in Virginia Beach, mid-sixties. It was on its first controlled deep dive, and they don't — I don't remember all the details. A couple of students over the years have chosen it for their presentation. They had a hard time even recovering parts of it. One theory is it was a bad braze joint in one of the seawater piping penetrations through the main pressure hull — four- or six-inch pipe just brazed. The Navy in the SUBSAFE program redesigned that. Rickover was still around, still in charge, and they shut down submarine construction for about a year and a half to three years. A lot of the manufacturing management techniques used commercially in the 1990s, the US Navy developed them under Rickover in the 1960s because of the Thresher.
They think there was a major penetration into the reactor compartment, they lost power. They didn't have the pumps to blow out the ballast tanks enough, and they were on a slope coming up, because they had a tender right above them that knew what was going on. They could hear from her sonar — these guys screaming inside the sub. They were on a slope toward the surface, but they hadn't been able to blow enough water out of the ballast tanks. They went up and they came back down.
Student: So the issue — over thirty percent of these joints were failing inspection tests. And if a seawater connected system fails, you spray onto the breakers that control your reactor — reactor scram. So you've lost propulsion. The ship initiates procedures for coming shallow: put up angle, use the momentum. But they tried to actuate an emergency main ballast tank blow, and at the time they didn't have the same kind of air controls we have for our emergency systems now. They suspect there was moisture in the high-pressure air, and a small pipe meeting froze with moisture inside — the pipes froze, couldn't actuate the blow. So at angle, the result was failure to initiate emergency main ballast tank blow. The means of their communication with the surface ship — they lost contact. The sub essentially slipped back down. It was trying to go up, but slips and plummets directly downward, past test depth, breaks into two pieces.
That's more detail than I'd ever heard before. For a long time some of that was classified — until probably the 1980s it was really hush-hush. The part that was hidden was to protect the Navy from embarrassment. But in fact the welds were terrible. It wouldn't surprise me if that's one of the reasons they eventually shut that shipyard down. Terrible quality control. A lot of the quality management systems in place throughout the United States — all the stuff with the Japanese and their quality control techniques — came out of the SUBSAFE program. Rickover was leading that, he had a blank check from Congress, and he didn't want to be embarrassed again. It was tragic. One of the tragedies is the people on the tender could hear it. From sonar records they think what they really heard was the sounds of the sub collapsing.
I personally heard that recording, given mostly to the command. You can actually hear the captain narrating as the pressure hull was collapsing. I heard it in a Navy annual training video.