§1. Schedule note and the Blackhawk washer follow-up [00:04]
Next Tuesday is actually Monday at MIT — they do this once every spring term so the schedule gets the right number of Monday-Wednesday-Friday and Tuesday-Thursday classes, and the way they do it in the spring, they always have to turn a Tuesday into a Monday. So next week I'll be lecturing Tuesday and Wednesday in this room at 9 o'clock. If you go to 26-100 at 11:00, have fun. The videos are going up within a day or two now; we had some problems earlier this week but I think it's all resolved. Simone will be doing both Thursday and Friday of next week.
We talked about the Blackhawk washer two days ago, and I brought a couple of my little pieces left over from fifteen years ago. [Tom produces the washers.] This is the washer, and fifteen years ago it cost $350. This one's painted gray — that means it's a Seahawk. This one's green, it's an army part. This one's steel, this one's aluminum, don't ask me why. It's a very clever design. In order to keep it tight, there's a different number of holes in this than in here, and so no matter how you turn it, within about two or three degrees you always get one hole to line up, and they put a screw in there to keep it from loosening. They don't use lock wire or friction; they have a positive thing, because if you lose your tail rotor you just spin to the ground, which is exactly what happened.
But the reason they lost their tail rotor is this washer. In the incident, one developed a stress corrosion crack. We took this one which had never been used — the reason there's a saw cut there, we measured the distance very precisely, within a fraction of a thousandth of an inch. There are flats here that keep it from rotating. Then we made the saw cut and measured the distance again, and depending on whether it shrinks or opens you can calculate — it's basically just a simple circular beam — how much stress it took to open it or close it. We measured 8 ksi residual stresses in this part. With the heat treatment for the alloy it's made out of, 5 ksi would cause stress corrosion cracking. The solution was to go to an overaged T73 heat treatment, which has a stress corrosion cracking minimum of 25 ksi. So they had been making it out of something that was only good for 5 ksi resistance; they switched to 25 ksi material, which is just a heat treatment change, not a compositional change. The residual stresses were 8 ksi, so everything fit together.
§2. Necking, voids, and inclusion control [03:38]
Now, a question — I wish I had brought Backofen with me, but fortunately they didn't erase my wonderful drawing over here. We were talking about necking and void formation in a tensile test, and these are graphs out of Backofen. You're doing a simple uniaxial tensile pull; here's the fractured specimen, here's necking. If you cross-section a piece of copper and do metallography, you find a bunch of voids forming in the neck. The neck goes into a state of triaxial stress — we'll talk about stress states later, but it's basically hydrostatic tension, and that's what helps the voids open up. The whole thing's being pulled in all three directions. At higher magnification you see these elongated voids, and you can see copper oxide particles — these little black spots are inclusions — and in the holes they're acting to nucleate something.
This, I think, is out of Hosford. It may be in your book. Hosford has some pictures and you can see the fracture in the neck, what we call ductile dimple rupture. This is a cross-section, and you can see the cross-section of the dimples and how the voids are forming, pulling apart. Classically, this is looking down on the fracture in the scanning electron microscope, and you see the ductile dimples like taffy pull at 3,000x. Inside many of these you'll see the oxide inclusions that are nucleating each dimple. That's classic ductile rupture. I've now preempted Dr. Belmar on ductile fracture.
Backofen has this drawing: if I had something with no inclusions, very clean material, and I pulled on it, it might neck down like this — I get tremendous elongation. As a thought experiment, the inclusions basically break the material up into a bunch of fibers, smaller tensile specimens, and when they neck down it's like necking down a bunch of fibers. The total elongation with inclusions is a lot less than the elongation if the material was very pure. We know that's true. Back in the '60s and '70s, we spent billions of dollars coming up with vacuum arc melting and all these fancy things to get rid of inclusions in metals, because they are crack starters.
All the crankshafts and piston engines and aircraft have to be triple vacuum arc melted now, because vacuum arc melting burns away the inclusions formed from the regular steelmaking process. You make the steel, cast it into an ingot, and then you take that great big ingot and arc it into another mold. You've got a round ingot full of dirt and inclusions, and you arc it into a copper mold, withdrawing as you go — you just remelt it, but in a vacuum. Now in the heat of those little spatters of metal going across, they're being melted in a vacuum, and it sucks out the oxygen and nitrogen, and you end up with a very clean steel. They will triple vacuum arc remelt aircraft landing gear, transmission shafts, and the shafts for turbine engines. They're using steel for these things and triple-melting them — once, twice, and three times. You end up with things so clean you look at them at 500x and have a hard time finding an inclusion. There are standards for all these inclusion counts.
§3. Orange peel and redundant deformation [08:24]
One of you asked me about orange peel. Here's a picture of orange peel — this is from Hosford's book too, probably chapter 16. I don't know what it is — looks like it might have been the bottom of a gas tank in the automotive industry — but it's a drawing from a stamping from sheet metal, and the surface looks like an orange. It's caused by large grain size. Those little pebble surfaces, each one is about the size of the grain in the material. If you want to avoid this — I'm not going to call it a defect, because it doesn't weaken the material, but if you wanted a nice fender on a car, the paint wouldn't cover it up. So it's not generally desirable. You have to control the grain size, and we'll talk about that later.
Going back several days, I was talking about redundant deformation and how you look at the deformation zone geometry. At a delta of one — sheet rolling, straight-line deformation — you have x's in the slip-line field. But at a delta of four or ten, the slip lines all have this curvature. When the slip lines are straight, you have uniform, homogeneous deformation. When they curve at higher delta, you have non-uniform, inhomogeneous straining — redundant deformation. The redundant deformation comes from the fact that you have to bend it one way, then another, then another. I could tell some of the eyes were glazing over at that point.
So here's my example. [Tom produces two decks of cards held together with rubber bands.] If I deform it, I shear. Remember, plastic deformation only occurs in shear. Hydrostatic compression or hydrostatic tension has no effect on deforming the metal. It might allow it to go to higher elongation, but it doesn't cause the shape change. The shape change is due to shear. So if I do a shear on this, that's a delta-equals-one type of shear. At delta equals five, you shear it that way, you shear it back the other way, and then you shear it that way to get the same shape change. You did three times the work to get to the same thermodynamic state. In thermodynamics, the end states are state properties — they tell you the energy or the shape, but they don't tell you the path. Work is not path-independent; work is not conserved. The shear work to go this way, back that way, and then this way again — I get one shape change for three, but I pay the price three times in energy. That's redundant deformation. And a lot of that inhomogeneous flow comes from friction.
This is out of Backofen, and it shows the delta dependence of the inhomogeneity factor for copper strip with different lubrication. Here's delta. At delta equals one, the height and width of the deformation zone are equal — you're really in a rolling situation. Over here you're getting toward drawing or extrusion. You can change the die angle of the draw. Unlubricated, the inhomogeneity factor increases as you get to higher and higher delta. We saw that in the handout I just gave you — the inhomogeneity factor increases when you have less die angle. Then if you've got paraffin oil or soap — and remember, soaps are stearates, and later we'll talk about why stearates are such good lubricants.
We can also talk about why, if you live in an area with hard water and you take a shower, you still feel soapy after you've rinsed off. Anybody live in an area of hard water? You lived where? New Jersey. If you have a lot of calcium in your water, that's called hard water — the amount of calcium in the water is called the hardness. New England has nice soft water; take a shower, rinse off, you feel nice. But in San Antonio or certain parts of New Jersey they're probably using well water, and it gets saturated with limestone and a lot of calcium. The soap is a calcium stearate, and if your water is already saturated with calcium, you can't dissolve off the soap. A monolayer of stearate molecules stays on your skin, and you can rinse till the cows come home and you're still going to feel slimy. I travel around the country, go to a hotel, take my shower, and I can tell you whether it's hard or soft water.
§4. Upset forging and friction at the die [14:57]
Any questions? Now we want to talk about forging-type deformations. This is almost a delta of one, not exactly one but close. Not everybody uses delta — Backofen liked to use it, and his students Hosford and others like to use it, but before the 1960s most people don't talk about delta operations in terms of deformation. This is showing two different friction levels in metal flow. You start with something like this, reduce it 8%, 22%, 33%, 44%, and 60%, and you can see the inhomogeneous deformation. The squares in this rectangle turn into — in the center maybe they're still squares, but at the edges they have funny shapes. As you go further, you get worse and worse uniformity.
This shape change would be a frictionless type of deformation, but you typically have something that causes this barrel shape. If you want to see the barrel shape in the real world — this is Japan Steel Works, doing upset forging of some big heavy piece. Here's upset forging of a big hot piece of steel and you can see it's starting to barrel.
This thing's pretty big — it's a 600-ton forging. The largest forgings we do, or the largest pieces of steel we do, are about 750 tons, so they don't get much bigger than this. Anybody have an idea why it looks kind of whitish right underneath the clamp? Because we broke off the oxide. In order to heat this all the way through, you had to leave it in the furnace for a week — an air furnace at about 2200°F — and the iron oxidizes in the furnace. The oxide's about an inch or two thick after a week at 2200°F. Now you grab it with these big tweezers — they're big tweezers to pick up a 600-ton piece — and the oxide breaks off. The oxide has a different emissivity. When you actually forge this, you'll have this hot slag of oxide and it'll all break off.
I remember the first time I worked for Bethlehem Steel during the summer, they had a program to introduce all the 500 new engineers that year to steelmaking. In the morning we would listen to vice presidents give talks. This lasted for a week, which tells you how many vice presidents they had — you could spend five mornings listening to them. They weren't long talks, half an hour each. In the afternoon we'd go down to the Bethlehem plant. I remember going to the forging plant — this was summer of 1975 — and they had a big forging hammer slamming down on these things. When the great big crane moved the steel back and they were done, two bare-shirted steelworkers went out with their shovels scooping up all the slag, the oxide that had come off. It was about six inches deep on the forging bed, and you had to scoop it off, because if you brought another one in, all you'd do is forge that stuff into the surface of the next one.
It was all done manually, and it was amazing to me. I sat there kind of bug-eyed. They had a dumpster brought up, and one guy would scoop down, come up, and throw it over his shoulder into the dumpster — hot slag at about 1,000°F. The other guy would go down, and they were like pistons going back and forth. I said to a guy, what if one of them got out of sync and threw hot slag in the other guy's face? And the guy says, well, they've been doing it for thirty years. The guys who do throw it in the other guy's face, they probably don't work there anymore, because they may be dead.
The die starts out cold, and by the time you end up forging, the anvil — which may only be four feet thick — and the hammer above, which may be two or three feet thick, you don't want to touch it. It's probably pretty warm; it might get up to 1,000°. You're forging something at 2200, and you're hammering it. Go watch the blacksmith here in the basement. He starts with a cold hammer, hits something hard, and when he's done, that hammer gets warm. But it only gets to about half the temperature of the thing you're forging. If you keep hitting with that same hammer and don't let it cool off, you'll have a problem, because it heats up, and now you're taking one material that's softened by heat and another that's strong by low temperature, and you're using the hard cold material to form the hot material.
§5. Three forming regimes and isothermal forging [21:16]
That's something I was going to talk about next week. There are three ways we can form things. We either start with something that's really hard and form something really soft — remember those little machine screws, that's cold heading. That's all at room temperature, no temperature effect to give me a strength difference. What do I do there? I heat-treat the steel to 200 ksi strength, and the workpiece is going to be 30 ksi annealed steel, soft as I can get it. As I'm working, it gets up, and that final part is probably 100 ksi, but the tool is still 200 ksi. I've got to have good lubrication. So that's a case of same temperature but big strength differences.
Then I can use temperature as the advantage, but now the process isn't isothermal. I have to have a cold tool and a hot workpiece, and I have to have a little time in between for the tool to cool off. That's what you were asking — it gets hot.
The third one is called isothermal forging, and it was developed because of Backofen and superplasticity in 1965, when he rediscovered it from Germany. That's Tuesday's lecture; I don't think we'll get to that today. Isothermal forging is how we make a lot of the turbine discs for airplanes, jet engines and so on. That sample should be in the Smithsonian. I recovered it from a graduate student's desk — they were using it as an ashtray.
Student: Will they just let the forming tool — or do they try to cool it?
No, actually, in superplastic forming you get material that, instead of taking a 30% stretch, will take a 400% stretch. Sometimes you can use graphite dies, which aren't particularly strong, but you get the material up to where it's like clay. You take the nickel-base superalloy up to 2200°F and it will deform like clay. You have to have something that's a little stronger, but it can be something relatively weak, like graphite dies. You can use molybdenum dies, or other superalloys that have a little more temperature capability. You're limited in your die materials, but everything is hot — it's isothermal. The dies are just as hot as the workpiece, and you're using a material that's superplastic.
We'll talk about isothermal forging next time — I have to explain how you get superplasticity first. We could not build the jet engines you fly on today without it. When you fly on United or Delta or American, you're flying in a whole different technology than in 1960 when I flew up here to come to school as a freshman. They may have been jet engines, but they're a lot different.
§6. Nasmyth's steam hammer and stories about Backofen [24:44]
I want to show you one other thing. This is James Nasmyth's painting of his patented steam hammer forging the propeller shafts. This is a 30-inch diameter shaft for Brunel's SS Great Britain — it was going to be one of the largest ships. These were the largest wrought-iron forgings of their day. It was just a steam hammer, big steam cylinder up here, and they're going to take an ingot of wrought iron. This isn't even steel technology, this is wrought iron. Look at the manipulator they have back here — count them, there are actual men moving this thing. You've got an anvil right here. Notice the shape of that anvil — it's V-shaped. Remember, I showed you the picture of how you don't want to do deformation at delta of 5 with just two platens coming together, because you'll crack the inside — it goes into tension. But if you do it from three sides with a V-shaped anvil, you actually get compression in the center, and you don't cause splits in the center. 1839 — they had studied Backofen, they knew all about delta, and that's why they were using their V-shaped anvils.
Someday I have to tell you some stories on Al Backofen. He was an interesting guy. I'll tell you a couple of them right now. When I came back as a sophomore, I had worked 80-hour weeks that summer between freshman and sophomore year. It was during the Vietnam War, and I was lucky enough — I took the civil service test in Virginia Beach, where I grew up, to get a summer job. You had to be able to add two and three rather than just two and two — really hard questions — and I got above a 95, maybe a 98 or 99, which was high enough that I got hired as a postman.
The job I'd had the summer before paid me a dollar an hour as a clerk in a drugstore at Virginia Beach. The post office paid me $3.35 an hour, which was the standard rate. I was replacing postmen going on vacation for two weeks during the summer. My favorite part — later in the summer they had me delivering special delivery letters, like express mail today. This was before FedEx; for 35 cents you could get a letter delivered special delivery. Typically I'd show up at 9:00 in the morning, they'd give me ten or twelve special delivery letters that had to be delivered immediately in Virginia Beach. I'd get in the station wagon and drive for the next eight hours, because Virginia Beach at the time was the fifth-largest city in the United States by area. Chesapeake, Virginia, was the third-largest. They had incorporated whole counties under Virginia law into cities so that Norfolk and Portsmouth couldn't annex them. So they were paying me a little over $24 to deliver $3 worth of mail. And we wonder why the post office is going broke.
The next summer I needed to make my three bucks an hour to pay MIT tuition, so I did the civil service thing again. They hired me at the Naval Air Rework Facility, where they rebuilt jet engines — we had all these crashes coming back from Vietnam and had to rebuild the engines. I was just an engineering trainee, but that was only two bucks an hour, so I took my old clerk job back at a buck an hour, and I worked 80 hours a week at two jobs 30 miles apart. It was not a great summer.
So I came back here and decided I would work in the lab. I'd taken my mechanics of materials class — my first materials course other than 3.091 — Al Backofen was the professor. I went up to him at the end of class and said, Professor, would you have any jobs in your lab? He said, come back to me next week. Well, he came to me at the next class and said, I found a job for you, go see Tony Thompson. So I started working for Backofen. He was a very kind man as long as you weren't his doctoral student. If you were a bachelor's or master's student, he was a gracious gentleman — white hair, always wore a bow tie, kind of bow-tie academic, tweed sport coat. When I started, they put me in that little short door, 8-137 — that was Backofen's graduate students' office.
Backofen was leaving the week that I came. He had retired in his mid-fifties. He couldn't get along with anybody, decided to heck with it, and was leaving MIT just as I was starting. But he had liked me — I'd been one of his undergraduates — and he came to me and gave me a piece of advice: the only way to keep your sanity around here is with a pocket full of Valium. I've always remembered that.
He had a home up in Marblehead. This was 1976, and he and his wife decided they would put it on the market for an outrageous price: $90,000. If anyone offered it to them, they would just move to their farm in New Hampshire. Well, the first person who came by offered him $90,000. He obviously didn't have a clue what the place was worth, and he sold it and moved to New Hampshire. About twice a year he would come back when I was an assistant professor, and he'd always stop by to see me, ask how I was doing. In his later years — he only passed away about two years ago in his eighties — he had started a Christmas tree farm in New Hampshire. This was before you could buy nicely pruned Christmas trees. He'd go out in the spring and spend his time pruning the trees, and he'd sell them for $3 a foot — a 7-foot tree was $21. Ordinarily you'd pay $3 for a scrawny Christmas tree. He was leading the market. He was selling 10,000 trees a year and making about $50,000 or $60,000. My salary was $15,000 as an assistant professor.
The other thing he'd do up in New Hampshire and Vermont and Maine — he'd go to the yard sales and buy junk, then take it to New York City and sell it as antiques. One man's junk is another man's treasure. He told me he was making more money than he had ever made as a consultant or a professor. Backofen was a character. Maybe someone will tell a story about me being a character someday.
§7. Inhomogeneous flow in forging and extrusion [33:06]
I want to start showing you some types of flow you get from inhomogeneous deformation. Here's hot upsetting of steel with sticking friction. Sticking friction means the stuff under the tool doesn't move at all. The material is stuck to the workpiece — all you have to do is not put on the glass lubricant. For steel forging, like I showed you, you don't use any lubricant; you just let the oxide scale fall off. But for a nickel-based superalloy, you encase the whole thing in a glass frit, and when you get it hot, you have this syrupy layer of glass — the glass melts and gets thick, and if you pick the right glass for the forging temperature, it'll be a good lubricant.
This one was done without lubricant, and you can see the material flow lines. Initially, this material right underneath the tool didn't deform at all — it might as well have been part of the tool. You can think of the flow lines as grains of wood in the original billet of steel. It bowed out, and notice how it's curling around. You're going to form a defect — if you kept this going, you'd get a little lap. That's the effect of friction, and you can see the inhomogeneous deformation. This material didn't get worked at all; this material got worked in this direction, that direction, a number of times.
Let me give you another example. This probably comes out of Metal Forming by Altan. This is a titanium blank, which has an alpha-and-beta microstructure initially. It's been squeezed down, and from the etched microstructure afterwards, you can see it looks fairly uniform — little barreling at the ends, but it flowed fairly uniformly. This is an all-beta microstructure. Titanium comes in alpha form and beta form — alpha is hexagonal close-packed, beta is body-centered cubic. If you forge something in the alpha-beta range, good deal, nice uniform product. Try to form it in the all-beta — look what happens. You get shear bands, localized shearing. You've got the big coarse microstructure, and then these bands in between. You see the X from the delta-equals-one type of deformation. The material tried to shear on these simple little X-planes; the X got distorted, but you can see the X. Those are the shear bands.
This comes out of Hosford. One of the things people do is take modeling clay. I'm not kidding. Now they do a finite element analysis in the computer, but in the old days they would take modeling clay, put two pieces together with nice flat surfaces, scribe grids on them, and then extrude the two pieces through the same die. They'd pull it apart, and you could see the flow lines. Here we have different die angles: 40, 45, 60, another angle. This is actually Don Blickwede again, my old boss when he was vice president of Beth Steel. Those may actually be steel billets rather than clay — cold-extruded billets.
If you look at the hardness number — they're looking at hardness at different amounts of deformation, distance from the center. With zero deformation, everything's nice little squares and you have low hardness — if it's steel, let's say Rockwell B 50, very annealed steel. As you go up in deformation from 4 to 6 to 13 to 26, you'll see that initially you get very inhomogeneous deformation across there. Then you get a lot of deformation, and everything becomes homogeneous again because it's all fully kneaded. You've really beaten down that pizza dough. If you're going to make a pizza, it's going to have a uniform crust. These are the practical things in this course. You've got to knead it a reasonable amount. You can't do something in between. If you do nothing, you don't have much. The first stuff introduces a lot of inhomogeneity, and you really have to get to 40% deformation before things start to get homogeneous again.
I used to think, when I was a bachelor's student here doing wire drawing for my thesis to make superconducting wires, that I would do better with small reductions. I had a great big set of wire dies — you could buy diamond dies for 20 bucks apiece back then. I had a government research contract, so I went out and bought $1,000 worth of dies in 5% increments. That was the wrong way to go. You should be doing 30 or 40 or 50% draws, not 5 or 10% draws, because with 5 or 10% draws you get lousy deltas. High-delta operations let you crack the center, and you also get very inhomogeneous deformation. This is the highest delta operation among these four.
§8. The Acela trolley wire and the Phelps Dodge story [39:28]
If you're going to work it, you've got to work it a lot. I guess I do have another story on that. Here's the trolley wire again. [Tom holds up the trolley wire sample.] This trolley wire has got an equivalent delta of about 30. Super-fine grain — it's actually starting to get into the superplastic region, I believe. Made by Phelps Dodge.
The company that originally bid on this — this was for the Northeast extension of Amtrak, the Acela line from New Haven to Boston. There was a $330 million federal contract to build the line, and that was the product they wanted. They had to purchase somewhere between $4 million and $7 million worth of copper to mount it. I got a call about December 20th or 21st saying, Tom, can you go to Rome, New York? There's a little company making our trolley wire. This is from the Amtrak engineer. We drove out there on the 22nd, and they said, oh by the way, we need an answer before Christmas. That's not very far away from December 22nd.
This company had been making copper wire for years, but the largest diameter they had ever made was 5/16 of an inch. They had never made anything much above a quarter of an inch in diameter. The largest mold they had on their continuous caster was about an inch and three-eighths, and you can see the grains — there's a longitudinal cut and another cut. Huge grains. Remember I showed you the orange peel. If you have large grains and you deform something — Amtrak's question was, are we going to have problems with this large grain size? The company said, well, we're meeting your spec, except we don't have small grains, is that okay? I had to give the answer before Christmas.
On the 23rd I told them to reject the product — $4.5 million worth of product. It turns out it was a fixed-price contract with the US government, and the company had certified to the US government that this company in Rome, New York, that had never made wire bigger than 5/16 of an inch, was the only company in the entire world that could make this trolley wire — because Amtrak had some very tight specs. This is a copper-silver alloy that had to have certain mechanical properties, a certain amount of elongation. I didn't realize all this stuff, but I had rejected it, and Amtrak sent a letter saying we reject it, you can't put this up, you have to find other material. But they had already certified there was no one in the world that could make it to the Amtrak specs.
It so happens that Phelps Dodge, the second-largest copper company in the world, had been begging this company for two years to make a bid on this. The problem was, Phelps Dodge's alloy was so good and had so much redundant deformation and super-fine grain size that — Amtrak for some reason had a spec of between 8 and 14% elongation, and this stuff had 18% elongation. It had too much elongation. It was too ductile. It was too good. Ordinarily you have to meet a minimum, but they had given an upper limit. Why someone said the elongation should be between 8 and 14, I don't know. So this company hadn't allowed Phelps Dodge to bid because they exceeded the maximum, when there should have been no maximum. That's how they had written the spec, and technically they wouldn't let them bid, even though it was going to be a much better product. You've seen the tensile bars — I sent them around — they neck down, big stretch.
Phelps Dodge's price was going to be $6.5 million or $7 million, and this company in Rome was going to be $4.5 million. So the money in the pocket of this British company was going to be an extra $2.5 million of profit if they could convince the government to buy their junk. I'd rejected it, so it goes to a week-long arbitration in December at an ocean resort in Old Saybrook, Connecticut.
This thing goes on forever, but the bottom line was, when they gave the briefings, the attorneys for the British company argued that Tom Eagar was an absolutely insane and incompetent metallurgist to have rejected this material — he had no basis to do it, and it cost us $2.5 million and the government owes us $25 million extra because they forced us to go out and buy the Phelps Dodge material. The briefing from the government was: you committed fraud on the federal government. You told us there was no one else willing to bid on this in the entire world, and we found out that Phelps Dodge has been trying to bid on this. The attorneys from the British company realized they could be sued criminally. And in fact, that's exactly what happened a week after the arbitration. The FBI raided the offices of the British company, took the passports of the senior managers, took their computers. At that point I was out of it.
Don't defraud the federal government. Amtrak has the legal authority to put you in jail if you tell them a lie. Other companies tell lies to each other all the time, but it's not good to do it to the federal government — they can do things. The other part of that story is how we actually won the arbitration. There really was nothing in the literature that said you couldn't do this, that the grain size should be a big problem. But it turned out it was essentially an orange peel effect, a large grain size effect.
Even though Amtrak had told this company, don't you dare install any of that wire, they took about a 10-mile stretch down here in Providence and installed the stuff from Rome, New York. They ran the train along that 10 miles and videotaped it. The pantograph, which is the carbon brushes running along the wire — it turns out there was a rough surface on the wire, and as the thing's going along at 60 mph, it looked like the 4th of July. It was sparking all over everywhere. If they had put that up for a couple hundred miles, they would have had to take it down two weeks later, because it would have eaten itself to pieces from all the electric sparks. And the idiots were smart enough to videotape it for us. I didn't know at the time that was the reason to reject it. I did look at some data — Tony Thompson, who I first worked with under Backofen, had done his doctoral thesis on grain size effects of fatigue in copper, and I went back to some of Tony's papers. So that's another little story of how I got myself in trouble.
Sometimes you just have to make a decision. I wasn't sued directly, but Amtrak was sued for $25 million because of my decision, and we ended up proving ourselves right because they provided the evidence for us. I don't know what happened to the criminal proceeding.
§9. Extrusion dead zones and the Stamixco mixer puzzle [48:19]
This is extrusion. If you go with straight-edged extrusion dies, you can have different types of things going on, but in fact you end up with dead zones. This is what they call backward extrusion — you push a tube tool down into the billet, and up the center of the tube comes your extrusion. It's a question of what's moving: is your press stationary or is your press moving? But you end up with dead zones, and I want to show you some of the defects.
Here's one set of defects, or inhomogeneous deformation. This is a copper cast — you can see the grain structure from the original casting, and they extruded it partway. Pure copper over here; this is 70-30 brass done the same way. Look at the dead zones. Copper has a little dead zone — you can see the grain refinement. This is a reduction ratio of about 85% on both. But look at the shear bands in the brass. The different work hardening between copper and brass gives you a very different pattern and these dead zones. That kind of pure shear not only raises the forces necessary to do the extrusion, but it ends up giving you lousy surface finish coming out of the die, because you're rubbing metal against metal — galling, if you will.
The last thing we'll do is talk about Stamixco, which we were talking about with some people before class. If you're doing polymer extrusion — an injection molding press — it's like a great big toothpaste tube. It might be 30 feet long, made out of steel three inches thick, but it's just a toothpaste tube. You're going to squirt polymer out and extrude it into something. Part of the problem is, a lot of times it's like a two-part epoxy — you want to mix two polymers together. If they're very thick, it's hard to get them mixed well. So this is what they call a static mixer. That's where the name Stamixco comes from. Pretty simple tool, but there's some technology in the design.
You basically have this type of thing that goes into your extruder barrel — goes inline. You've got your pressure ram, your dies and everything up here. The polymer melt, or polymer in a partially molten state, is brought in here, you put it through all this, and you get it to mix. They've taken a blue plastic and a white plastic, brought them in — blue down the center and white on the outside — and put them through these elements. There's a three-dimensional model, and in just a few lengths you've got light blue: it's mixed. If you tried to do that without something to create the turbulence, it would take you half a mile with things of this viscosity. So this is extrusion of viscous polymers. Here's the starting material — blue core, outside tube.
Turns out the guy who runs this company is an MIT chemical engineer, has a PhD, very nice guy. He came to me once — we had done some work on something else — and he'd gotten together with these guys. He'd run a contest at a polymer extrusion conference: here's an extrusion we did with the blue resin in the center, and when we're all done, the blue resin is on the outside. It's not on the inside. How did the blue resin get to the outside? He had a $500 contest for the students. He wanted to run a contest at MIT to see if MIT students could do it, because out of 50 students at the conference, only one young woman figured it out. He was so proud of his thing.
So does anybody have an idea how the blue resin got to the outside?
Student: Diffusion?
Nope, not diffusion. What have we been talking about? Inhomogeneous flow. If you had looked at this thing at some step in between, you'd have found that you had white sticking by friction around the boundary on the wall of the mold, and blue bulging in front. So here's white, and here's blue. Keep doing that for a while, and the nose goes up further and further, until finally the nose folds over on itself and turns 180°. Look at the top here — it folded over up there. If you had seen it at an intermediate stage, you would have seen this stuff bulging — inhomogeneous flow.
He was very disappointed to find that I could solve the problem in five minutes. I didn't give you five minutes because we didn't have five minutes. It's partly because I think in terms of deforming metals. Here's a polymer problem. He's a chemical engineer. In his mind — and if you read some of the stuff, he explains this in terms of a boundary layer, because a chemical engineer or a mechanical engineer understands boundary layers and they think of the world in terms of boundary layers.
I had a problem once which was a problem in current flow through copper and steel when we were doing welding. The steel would go past the Curie temperature and become non-ferromagnetic, and that's going to really affect the way the current flows through the thing. I went to see Jim Melcher, one of the great scientists at MIT, who passed away about 15 years ago. We started talking diffusion. To me, diffusion is: you have a vacancy, an atom moves into the vacancy, and they change places. Jim Melcher started talking diffusion, and he drew me an infinite series of parallel resistors. He thought of everything as an electric circuit. I thought of diffusion in terms of atomistics. Mike Mousakas thinks of the flow in Stamixco as a boundary layer problem. I think of it as inhomogeneous deformation. Even though we're all engineers, we speak different languages. That's one of the lessons for today. I'll see you on Tuesday.