FW_Su2013_06

And then we'll get into corrosion as far as that goes. Okay, let's do this. So the eighth thing is Electron Beam is distorted by magnetic fields. Uh, in fact that can be a good thing. In fact we use magnetic fields to Traverse the beam across things. Many times — I had a student who was looking at solid Electron Beam welding of the steels, actually was looking at cooling rates of stainless steels, and he chose to use Electron Beam because he could get seven orders of magnitude of cooling rate. And um, the fastest cooling rate, he put a high power beam and he swept it across magnetically just across the plate, and he got a weld that was like 1,000 [microns] deep but it was cooling at about 10^6 degrees Centigrade per second because of all the metal around it and the small size of it. And then he had bigger welds and stuff. The challenge with the smallest one was just to be able to measure the samples that were that small.

Um, in fact magnetic fields are used to an advantage when you're doing Electron Beam Heat Treating. We talked about not necessarily welding but Heat Treating. Um, we talked about Electron Beam heating of the camshafts and things. Uh, typically people might use a raster pattern where you just, you know, like the old cathode ray, x-ray, um, x-ray cathode ray tubes on the TVs, where you just scan back and forth, okay. The problem with that though is you always end up with extra heat at the beginning and the end of what you're doing as you're changing direction, okay. You have to decelerate and accelerate. So what people have developed is Lissajous patterns. Anyone know what a Lissajous pattern is? No electrical engineers in here. If you take a sine wave and you use it as an input to the y-axis, and you take a different frequency sine wave and use it for the x-axis, you can get a pattern of swirls which is known as a Lissajous pattern. If you look it up on Google. But basically if I wanted to do a rectangle like this, if I scanned back and forth I would get too much heat on the ends. But with the Lissajous pattern you basically — well I can't draw it very well — but you get a bunch of swirls like this and the thing is never stopping and changing direction. And you end up covering the whole surface, uh, average time average, and you don't get those sharp ends of heating.

So far as that goes, I might even have a sample of a problem of Lissajous heating. I guess I didn't bring it. Anyway, um, or non-Lissajous heating. Anyway, so magnetic fields can be used to your advantage but they can also be used to your disadvantage. It turns out, uh, when the Japanese were doing these um very heavy section, like 6 to 8 inch thick Electron Beam welds for these pressure vessels — the gas, coal gasification pressure vessels — they found when they were welding dissimilar Steels, you actually would set up a thermocouple effect between the two different metals. I mean, you know, if you have two different metals in a temperature gradient you actually have a voltage between the metals. It's called the Seebeck effect. Uh, and because of the different thermochemical equilibrium for the compositions you set up a small voltage, it might be millivolts. But when you're doing these thick section steels you actually would get a current flowing from one steel to the other during the welding process. And that current has a magnetic field, and it would cause a straight beam to curve, and you would end up missing part of your joint if you were more than 4 inches thick. The thing would just bend right out of the way. And no one ever found a solution to that. So you couldn't weld dissimilar steels, heavy section steels with the electron beams because the beam is distorted by magnetic fields.

The advantage of lasers: nothing other than astrophysical, uh, gravitational fields of galaxies and things like that distorts the light, okay. So we don't usually have to worry about that, okay. Um, Electron Beam — we've already talked about anything above 60 kV, it's going to produce fairly hard x-rays. Um, and in fact, um, we either go up to 60 kV or we go all the way up to 100 kV or 200 kV. We kind of don't find very many machines in the 60 to 100 kV range, okay.

Um, turns out — I mentioned that — when as soon as someone gets a new heat source they tend to try to use it for — um, I'll pass the small one around, show you the big one — um, they try to use it for welding or something. And I told you about this situation where the Navy had spent a quarter billion dollars up to 1992 trying to develop these relativistic Electron Beam weapons so it would replace the Phalanx system. And you'd have this relativistic pulsed Electron Beam, and if someone has, you know, firing some missile at you, you'd shoot it through the air 30 miles away, you'd hit the target and blow it out of the air before it hits your ship. Well, peace broke out with the former Soviet Union. They were trying to figure out what to do. And they said — I told you they had a little workshop and they asked me, could you weld submarines? I said, well you could melt a submarine with 5 megawatts of power but you can't weld, because you only need at most typically 100 kilowatts of power.

So anyway, since they were trying to find a use for their technology, we had a research contract and we tried Electron Beam deposition. And this is a piece of nickel manganese, um, bronze, nickel manganese aluminum bronze, I can't remember. Uh, anybody know what this alloy is used for in the Navy? Propellers, okay. So obviously propellers are circular and you could potentially build up structure. So this is a piece cut from that, and we built up this Ridge with an electron beam. Uh, so far as that goes, this is a piece of stainless steel where we were building up uh a ridge just going around and around. But it turns out sometimes you end up getting solidification defects, surface tension defects, and they tend to propagate and it's hard to get rid of them. The other problem with this technology is you develop residual stresses. As you build up bigger and bigger things, you have these small little weld zones and you put one on top of another and the of another on top of another, and the residual stresses get worse and worse. So typically you get to about more than about a half an inch and you find that you uh you end up cracking.

So 20 years ago we were trying to do solid free form fabrication. But it turns out, because of — I talked about the heat flow equation, it goes as the distance goes as a square root of time. If you try to go — if you go slow, then you're wasting a lot of heat, you're not getting efficiency. If you try to go fast, it turns out the faster you go the less depth you get of penetration, and so you end up putting down less and less metal because you're just doing surface heating, okay. And so whether it's laser, Electron Beam deposition, uh for solid — what they call solid free form fabrication, just feeding wire in and trying to build up — today it's also the big news in the Wall Street Journal, this is additive manufacturing. But people have been doing it for 40 years or longer, of laying metal down. Okay, over weld on top of weld. You've probably seen it on rebuilding of shafts in the shipyard, okay. You come in and the tail shaft is worn where the bearings are, and so someone comes in and machines it away and then they put some weld metal in there over and over.

What did you also see at the same time? A bunch of torches, flames on there, as they're heating it to preheat it and post-heat it and to relieve the residual stresses. Cause you put on more than about a quarter of an inch and you're going to have tremendous residual stresses which will cause the thing, the bearing, to fail in service unless you relieve the residual stresses. So you can repair things and build them up but you have to worry about the residual stresses.

One of the problems with these multi-megavolt uh pulsed energy weapons that they were trying to use is, first of all, when you start getting to a million electron volts, you're now talking about radiation shielding that might be 2 feet thick concrete, high density concrete, with lots of barium and uh oxide and stuff in there to shield the radiation. I mean this radiation will go through, you know — well, that's what we use for doing x-rays of 3, 4 inch thick steel. So it'll go through 3 or 4 inch thick steel. You need feet of protection. The other problem is, you get above 5 megavolts and you get to what they call the activation limit. And the activation limit is when those electrons coming in are sufficient to transmutate the um atom into another type of atom, like iron can go to Cobalt and in fact become radioactive in the process. So you can produce radioactive elements above 5 [megavolts]. So I mean the Navy in their particle beam weapon program built 45 MeV machines like out at Lawrence Livermore, and they were buried like 20 feet in the ground, okay, just for shielding. So I guess — but, you know, all you're going to do out there in the ocean is radiate a bunch of fish, right. Hopefully the Sailors are far enough away. Anyway, um, so there are problems with going to really high voltages in the x-rays that you produce in these things.

And people have looked at um making, building things back in their um, like 1976 or 77, when they first had some high power lasers. They tried to build up things, and I remember there was a research contract at Pratt and Whitney from The Office of Naval Research to use this sort of — not classified but this very proprietary, uh not secret but almost secret — uh laser, high power laser, to build up things like we were doing with electron beams. Um, and to try to see if you can make turbine disc, okay. Turbine disc could be worth $100,000 a piece. Uh, and if you could build them up — well the problem was you're laying down such thin layers and you have residual stresses. They were laying down 15 — or 15 pounds a day or something, which is not very much, okay. When you start talking about trying to build a thousand pound turbine disc or something that you're going to machine to a smaller weight. In any case, uh, people are still trying though. I mean after, you know, after 35 years, the people who tried it before and found it didn't work, they're going to — there's a whole new generation of people who refuse to believe that the people before them were not completely stupid, okay. Um, they believe they were stupid and they can make it work. So today additive manufacturing is one of the buzzwords. Um, and people are trying to do what we were trying to do 20 years ago with electron beams.

We thought with electron beams we had a — we had an edge over lasers, because with a million volt Electron Beam you can deposit the heat a mill, a millimeter beneath the surface. With lasers, with surface heating, all you can do is lay down at very high speeds these very thin, like, you know, 50 to 100 micron layers, and building that up is really slow. If you could lay down millimeter layers, we showed you could get 500 pounds a day. And so we thought we could build valve bodies and other things.

Anybody know why you need to build great big valve bodies but in a rapid way? I mean anyone done any procurement on ships, on a new ship design — on ships on a new ship design you might spend a hundred or a couple hundred million dollars on spare parts. Like great big castings, okay, or tail shafts or bearing housings or you know these 20-30 ton objects. Uh, because the lead time — anybody know what a typical lead time is? A 20 ton casting? At least a year, okay. So if you happen to have one crack, you got a big problem. Because you can't just go down to the Home Depot and buy another, okay. And if you go try to purchase one, it's going to — you'll say, yeah we'll make you one and we can deliver it in 70 weeks. Well, that's not a very good thing if your ship's, you know, incapacitated.

So they buy spare parts, okay. Great big castings, they'll buy a couple hundred million dollars of spare parts for the ship class. And they'll leave them maybe partially machined or unmachined, and then when a particular ship has a problem, they've got this casting sitting out in some field somewhere and they can machine it and turn it around within few weeks or a month or whatever, okay. But if you had some way that you could produce a new casting in uh a few weeks, then you could avoid all those uh procurement costs for spare parts when you're dealing with great big things. In fact I have a failure right now in a gas plant — has in uh in Wyoming, uh they had an explosion and they lost a critical heat exchanger. And this heat exchanger, the lead time is 70 weeks to buy a new one. So you got a quarter billion dollar project that's down for at least 70 weeks unless they can find a spare somewhere. Well there's not a lot of spares of these sitting around doing nothing, okay. $20 million heat exchangers, aluminum heat exchangers for cryogenics, they just don't exist. Um, well, they exist but there — someone else is using them, okay.

With lasers the problem is you have uh invisible infrared reflections around which create problems for you. Um, and that people have to wear glasses, uh you can't see where things are, um where the beam is and whatnot. So you have to be careful. You go into a shop that's doing um laser cutting of plastic or cutting of steel or something with some robot, the whole thing is surrounded by, you know, a fence essentially with inter-locks and stuff, so no one walks in to where the robot is, because it's not a good day if they put their hand through the beam, okay. Um, so you've got to worry about the infrared reflections. Um, actually you have to worry about the visible light reflections, but most of them are infrared, uh most of the laser, the high power lasers.

Um, currents limited to about a half an amp — I think we talked about that. In practice we use about a tenth of an amp, because the electrons are close enough together above at a half an amp that they just start defocusing the beam. The laser on the other hand, you'd like to have what we call a nice Gaussian beam, but in fact you don't. You get a nice Gaussian beam — I was — this is not exactly correct what I'm going to show you right now but it's in principle the correct idea. If I had a laser spot, and in one direction I have a half sine wave, in another direction I have a half sine wave, I would have areas where it's uh additive and areas where it's uh negative, and this would be what we call a nice Gaussian beam. But if I have, uh — I'm sorry this was not positive and negative, this should be essentially a nice positive right in the middle of the whole thing. If I have one that goes with a full sine wave in this direction and a half sine wave in this direction, I'll have positive, negative. And if I get multiple reflections of sine waves I'll get fancy spots.

This happens to be one of the things they do in the laser business. They just take a piece of plexiglass and they pulse it with the beam and uh just vaporize some of the plexiglass, and so you can see what your spot size looks like. So this came from a laser which had a cavity that produced about 7/8 inch diameter beam, and you can see all the little diffraction spots on here of positive and negative reflection. They also can tilt that beam, and down here if you look through it this way you can see they got a higher heat intensity for — they were going to do laser uh surface heat treatment with a linear beam. And you get that by taking a circle and just turning it sideways, and the projection of the circle now becomes an ellipse heat source. So you can get kind of a linear heat source and you can do surface heating of things.

Um, so lasers have diffraction spots. If you want a nice Gaussian beam you lose a lot of your efficiency — over half your efficiency is gone. The laser mode cavity prefers to have all these unstable modes, or — I don't know if I should call them unstable — these multiple modes that degrade the beam, at least in terms of welding where you'd like to have a nice uniform kind of peak Gaussian heat source.

Uh, so I guess I said it's affected by magnetic fields, you can manipulate it with magnetic fields, and with lasers you use mirrors or fiber optics. One of the advantages, um, in recent years for laser heat sources is the fiber optics have gotten better and better. In fact Professor Yoel Fink [Joannopoulos? — see register] of this department did his doctoral thesis developing the, basically, a perfect mirror, okay. And the perfect mirror, you put different layers of different index of refraction materials and you can get a quantum effect such that the light cannot go through, and so it has to reflect back. And so the first big application of this, when he was a young professor 15 years ago, 20 years ago, 15 — yeah, about 15 years ago — was to make fiber optics for laser surgery, because high value added medical business. But now people are using some of these things for higher power lasers such as laser welding. Uh, and so you can use fiber optics to get the beam there. In the old days you had to use mirrors. And I already told you the problem the other day about mirrors and you get any dust on the mirror, and when the laser beam hits it, it just couples very well to the dust because the dust is an insulator, it's not reflective, and you just burn a pit right in the surface of your mirror. So you can destroy some very, equipment, very expensive Optics very quickly.

Um, the last thing is electron beams heat the sample uniformly, okay. Everything gets heated the same. Partly because the electrons have a certain depth of penetration. From 100 kilovolts it might be a tenth of a millimeter, at a megavolt it might be a millimeter, uh penetration depth. So it averages the heat out over inclusions and other things. Whereas it turns out the inclusions in steels are oxides, and they couple with the laser energy better than the base material. So 30 years ago the Naval Research Lab had a 25 kilowatt laser to study laser welding, and they found they were welding HY-80, and they got better properties in the weld metal of the HY-80 than they did in the base materials. So mechanical properties were better. And they said, well why is that? So they went in and looked. It turns out they were refining the steel. The laser was coming in and it was essentially selectively vaporizing the oxide inclusions in the steel, oxides and sulfides, and just blasting them away as vapor and leaving behind a pure, more refined weld metal. So there are certain advantages you can get in changing the chemistry of the weld bead so far as that goes.

Student: Is that a practical way to apply that to producing steel stock in the first place?

Well the only problem is your melting rate. Uh, with lasers, you know, there are megawatt lasers but they're really pulse lasers and they're chemical weapons and they have spot sizes like this, okay, cause you're going to try to shoot down a missile. Uh, if you got a laser that you want to focus, we're limited to like 25 kilowatts. 25 kilowatts, we're probably talking 50 pounds an hour, so forget it. You know if you start working out the economics of the production rate it's just, it's not practical. And then if you start looking at trying to build these things up in layers like we did those Electron Beam layers, you got to have something that actually deposits the heat beneath the surface. The laser just heats on the surface. And the faster you go, the thinner the layer. And it's a decreasing returns because of the fundamentals of transient heat flow where x is square root of alpha t. Okay, the faster you go the thinner x becomes, and x becomes thinner faster than time of melting increases. So you just get thinner and thinner layers and the net melting gets to be less and less with time, okay, as you go faster and faster. So there's some fundamental limitations to these things unless you can get the heat beneath the surface.

And we actually showed with these higher, these megavolt electron beams, you could deposit a millimeter deep. And we did a model showing the traveling Gaussian heat source with heat on the surface and heat beneath the surface, and showed you could get 500 pounds an hour, and I estimated it cost you $10 million to build a facility, but you'd have to bury it in the ground about 20 feet and it'd have to be completely automated, no people around, because you could give them a lifetime radiation dose in, you know, 30 seconds or something. Um, so there are problems with some of these things. Couldn't convince anybody that they wanted to build it, okay. But, you know, there are people out there — you hit the right person at the right time who has the — who's just experienced the most expensive problem that he can imagine, and he'll say, okay I'll spend $10 million to build that. Um, and it still may not work because there's lots of problems to be worked out. Once you, even if you develop something, the development problems get to be horrendous.

Um, so — which is another thing — if you're entrepreneurs, you know that of every 20 good ideas only 5% work. One out of 20 actually works and becomes commercialized. That's across the board. Um, so what looks like a good idea at the research stage has a 5% chance of being a good idea in application, okay. And that's what — well I shouldn't say that's what entrepreneurs are about because that is sort of what entrepreneurs — it's not what venture capitalists are about. Venture capitalists basically, they will look, they will sort through a hundred different possibilities and pick one. Which one out of a hundred which they will choose to invest in. And usually it's already a success, a proven success. And they pat themselves on the chest, say weren't we smart to pick this technology when it was obvious — it was actually obvious, anyway. Um, but that's another story.

Um, most of the people who think they're great businessmen are — um, it's just kind of what they have to, either the position they're put in or who they have to compete with. I mean, the, uh, you know, I've been watching over 30 years, and American businessmen like to say they're the best managers in the world. But you walk into American factory and you say, this is the best in the world? I mean you know you look at some of the management practices, and then you realize that, uh, back 25 years ago we were competing with Japan, and Japan actually looked like they were beating us. But they — we actually had higher productivity than the Japanese. But the Japanese actually have built into their economy a 30 or 40% lack of efficiency.

Anybody live in Japan for a while or anything? You know what the lack of efficiencies are in Japan? It's coming back to bite them now. But when I went over in the mid 80s, everybody thought, oh, Japan rules the world, because they were producing the best cars and shipping them to the United States and stuff. Well, it turns out the Japanese have this culture that they want to be an agricultural society. So 30% of their people are in agriculture. Uh, do anybody have any idea how many people in the United States are in agriculture? About two [percent] now. It's been — back in 1790s when the country first started we had 95% of the population was in agriculture. And over the years, as you know, agriculture has gotten mechanized and fertilizers and all these other things, we now can feed the world with about 2% of the workforce.

Well, Japan still has people out there planting rice by hand and they have 30% of their workforce in agriculture. How do they support that? I — when I was in Japan I figured a great gift — the Japanese like to give gifts — a great gift would be a pound of wild rice. I mean, Japanese called, um, well they call America, okay, but the — one of the symbols they use for America if you look at the Chinese characters is Beikoku, which means rice country, okay. Uh, because we are agricultural nation and they respect that. But I was told I could not — if I, if Customs would have a fit if I brought a single grain of rice in, because the Japanese have all kinds of things to protect their farmers. They only need 3% of their workforce in farming but they have 30% of the workforce in farming, so they have an inherent inefficiency in their economy. And so they complain about their poor economy. But if they had a free market economy and got rid of all these rice farmers and found something for them to do, then they would have a much better economy.

On the other hand, American businessmen — so that's an advantage American businessmen have competing with Japan. What's the advantage in Europe, particularly France? In France, if you, if a company hires someone, you own them for life. You can never downsize or anything else. It's sort like GM used to be. The unions in France are extremely strong, and if you follow the papers you know the French want to keep their 20 days a year of vacation and they, you know, they like their 20 hour work weeks and things like that. And so that's what we're competing with in Europe is an economy where the businessmen have just — anyway — who got, are strangled by their own regulations in their countries. So one of the advant— and so American man, American businessmen say, why aren't we great businessmen, we're the most productive in the world and we do all these wonderful things, and you walk in their plants and they're doing stupid things, but it's only because they're competing with people who have inherent millstones around their neck, okay.

Um, heavy section Electron Beam. So I spent a year in Japan in the mid 80s cause the Japanese had spent a couple hundred million dollars on a national project to try to build these coal gasifiers and weld heavy section steels. And someone's going to come along when you're a captain or something in the Navy and say, oh, we're going to build an electron beam machine for you, can weld um 2 inch thick steel very rapidly. Well, what they found was long-term beam stability. If you're going to weld one of these 30 foot diameter pressure vessels 6 or 8 inches thick, it takes about an hour to go around, and typically you'll have some hiccup in the beam every 10 or 15 minutes, which means you're going to produce a defect every 10 or 15 minutes. And when you produce a defect, as you're welding continuously, um, you now have to worry about how do you repair those defects? Well I've got this narrow little weld, okay.

The advantage of laser welding is we have something that has a depth-width ratio of about 10, okay. This was your original seam, and you got to try to first find it. But also that's another thing on here is non-destructive testing is not very sensitive. But if you find a defect, you got to come in and cut it out and reweld it, and how do you do that, okay. Uh, there's not a good technique for repair of deep seated defects inside something that's very thick like this. Non-uniform penetration — it turns out I talked about, in the vertical position, um, how the metal would slump down. In, you have a big deep cavity, it's like you're almost drilling a hole and the metal's back filling it. Well every now and then it back fills and you'll get a spike in your penetration. So the penetration depth is oscillating down here at the bottom. And they never found a solution to that other than welding in the horizontal position which worked for certain things, okay.

High equipment cost — you might be talking a 10 or 12 million dollar machine. Um, low utilization factor, it might weld for 2% of the time during a week or less. It required extremely clean steel. The same um situation that I talked about, the laser would preferentially couple to the inclusions — well it turns out the Electron Beam is vaporizing all the metal there, getting it to a very high temperature. And any oxides or nitrogen or anything else would come back in the steel and create porosity or um poor fracture toughness. And it turns out they had to get steel from the steel mills which was exceedingly clean. I mean we could make it but it was doubling the cost of the steel in order to get sound welds, it wasn't full of porosity and other problems.

Seam tracking is critical. Well it turns out you can make welds — the Welding Institute has made welds, Electron Beam welds in aluminum with the depth to ratio, depth to width ratio of 50 to 100. But if you're talking about a 2-inch thick weld with a depth to width ratio of 50, then you're talking about a width it's only 4,000ths [of an inch] wide. And that means if you're off by one human hair side to side, you've missed your joint, okay. Or if you're off angularly less than one fraction of one degree, you missed your joint, because you're skewed at some angle to your actual intersection. So it turns out in practice, people like to talk about huge depth to width ratios in welding with lasers and Electron Beam, but in practice you always have to be less than about 10 because of seam tracking, uh tracking the beam to be right where the seam is.

Uh, fit up is another problem, you got to machine these things if you're going to be that precise. Poor fracture toughness they found in many of the steels. Um, the non-destructive testing — typically if we've got, you have to take this, one of the modules you have to take has some non-destructive testing, and probably somewhere in there I'll talk about the fact that most of your non-destructive test techniques can see something about 2,000ths, 2% of the thickness. So if you have 2% sensitivity on something that's 8 inches thick, the smallest defect you're going to find is about 3/16 of an inch, okay. So sometimes that's good enough, but it's uh not always sensitive enough.

Uh, stop start locations in craters are essentially — if you're going in a circular uh pattern to make a pressure vessel or something, um, you're going to have to start and stop it at some point. And then your end crater as you power the beam down and let it solidify, you've got to do this in a controlled way. Um, and so uh they never completely solved those problems. It essentially became a defect that had to be repaired at the end and they never got, solved that problem. You've got to have a very large vacuum chamber, or local vacuum with all the problems of local vacuums and sliding seals and things like that.

Um, and it turns out as you get thicker and thicker, the range of process variables gets less and less. For example with a 20 mm thick steel, I've got some notes here that shows that the power that you can use varies by about 30 or 40% and you'll still get good welds. But if you go to a 80 mm thick, four times as thick — 80 millimeters being a little over 3 inches — you have like a 5% process range, not a 40% process range on your power to get a good weld. So your process variables narrow down to next to nothing.

So, um, anybody have any questions? Welding with lasers and Electron Beam — the theme is they're actually very different processes. The thing is, you can get very thick sections in theory but you have all kinds of big problems um that people haven't really solved. So most laser, Electron Beam is not done over about a half an inch thick, okay. So you know people have been saying this is the solution to heavy section welding, but no one's ever — I, well the one application I know of, I think it was Mitsubishi Heavy Industries, had to make a bunch of spherical stainless steel vessels that were 2 inches thick. And they built a chamber that was about five time — well it wasn't five times, it's probably about — the vacuum chamber was about two or three times the size of this room. And they could fit the whole vessel in. It's 2 inches thick stainless steel and they could go around it in about 10 minutes. And they had to make hundreds of these of some nuclear reactor or some nuclear physics experiment or something. They had to make hundreds of these little vessels and so they could justify this very expensive vacuum chamber that was the size of a two-story building, okay, to weld these things. But that's sort of a specialized thing. And stainless steels are easier to weld than carbon steels in many ways due to their metallurgy and whatnot.

So high power lasers and electron beams, very useful in the automotive industry, their greatest application is not for welding. Uh, but lasers in fact is just cutting, replacing sheet stamping operations. Um, lasers are becoming the norm in a sheet metal shop for cutting the steel — the cutting the steel, okay. But actually, and we do a fair amount of laser welding.

I — one of the problems, I guess I didn't mention it when we were talking about the reflectivity problem on aluminum. The reflectivity problem on aluminum is worse than it is on steel because aluminum reflects about 95% of the light coming in, because it's a very good electrical conductor. That means it's got lots of free electrons. Lots of free electrons means it's a very good mirror. So, um, people generally would say it's not practical to laser weld aluminum, except the exception that proves the rule is, at one time about 30 years ago, the largest application in the world for laser welding was laser welding of aluminum, in a plant in Illinois. Um, they were turning out 300 million feet a year of a bunch of sheet metal that had a cross section is like 10 or 15 thousandths sheet, and it had a shape formed like this to make a box. So this is the thickness is 10 or 15 thousandths, and they put a weld along here, okay. So this is the shape, and they had a weld right along here to seal up that seam. Uh, this was about 2 to 3 mm on this land.

Anybody have an idea of what these little channels of aluminum are used for? That's — two pane, double pane windows, okay. This was the frame. You put a piece of glass here — oops, should put something there — put a piece of glass here and put a piece of glass over here. You seal it. This is a separation. And they were just taking little pieces of sheet metal in aluminum and they would weld the seam. Now the thing, they were making 300 million feet, they had 10 lasers lined up in this factory just, you know, just pushing the stuff through like mad.

But the interesting thing was, because the aluminum has such high reflectivity, in order to — if you want to get a million watts per square cm in and you got 90% reflectivity, you've got to put 10^7 watts into your beam in order to get 10% in, right, cause you're reflecting 90, 95%. The problem is when you do that, this will drill a hole, this will weld. And you end up, when you actually couple, um, and you get a slight little surface depression in your aluminum — initially you have some melted aluminum on the surface, but then when that changes to melting through the surface, you now are depositing the beam down inside a cavity. And it, the light's re-radiated to the sidewalls and this little cavity absorbs the light much better than the flat surface. So you start oscillating between 10^6 and 10^7, between blowing a hole in the aluminum and actually melting the aluminum, okay.

So it turns out this weld was a very nice — if you look at a building from around 1980 — they use, actually, they use mechanical stitching now on most laser windows, uh double pane windows. But back in those days they were using laser um for the double pane window aluminum frames. And you look between the two frames on the windows and you would actually see a little weld, very uniform weld, looked like they made a series of spot welds. It was a continuous weld, it was just inherently unstable, and they were going so fast that it looked like a very uniform weld. And they wanted a weld that was slightly porous. You wanted this thing to breathe with the changes in atmospheric pressure. So they made a series of defective welds, inherently defective because of the properties of aluminum and the reflectivity of aluminum and the physics of the beam interaction.

Uh, so the largest application — the rule of thumb in the welding business is you can't laser weld aluminum reliably. But in fact the largest application in the world was laser welding of aluminum, and they were using it because it was so reliably unreliable, okay, if you will. It was just a natural oscillating process, you ran it continuous and it gave you a natural oscillation that was exactly what you wanted. That makes — they stopped because people developed the mechanical stitching technique, which is even faster and cheaper, you don't have to maintain lasers, okay. So another technology came in and supplanted it.

Okay, but that's sort of an interesting thing about — um, people will predict a new technology is going to take over an existing technology. I'm not talking about welding right now, just necessarily. And that's because they predict that this new technology will improve at a much faster rate than the old technology. But that assumes that the people with the old technology are just going to give up and say, okay we'll close down our business for you, rather than improving their technology to make it better. So when an old technology has no competition there's not a lot of incentive to improve it. But when a threat comes along, all of a sudden there is a threat. When there is a threat then people actually start getting creative rather than going out of work, okay.

So the example I like to use is beverage containers, okay. What did we use in the old days? Well we used clay pots. We don't use clay pots anymore, it's kind of slow to make those. But then we started using glass bottles several hundred years ago, right? You can still buy glass bottle beverage containers, right? Lots of beer bottles made out of glass. Uh, but then someone came along and started making um steel beverage containers. And then aluminum supplanted that, uh such that 40% of all the aluminum made in the world goes into beer cans and soda cans, okay, containers. Um, and then what was the next technology? Plastic, right. Go look right back there, beverage containers, right? They're made out of plastic. Uh, and which one dominates today? None of them. We got glass, we got steel, we got aluminum, we got plastic. And as one of them gets better and better, the other one gets a thinner wall, less material, whatever.

I mean if you found a — if you could find a beer can from 30 years ago, would have a lot thicker wall than that aluminum can today, okay. Um, they use supercomputers to design the shape of those cans, those soda cans, because they have a fairly high internal pressure. And in fact both the plastic bottles and the aluminum cans use the internal pressurization to give stability. You know that you uncork a 2 liter bottle of soda and you lose the pressure and all of a sudden you have a hard time holding it nowadays, it's so flimsy, right? And you have that problem with the water bottles now, you know you got to be environmentally sound, they got to make it thinner and thinner, such that now you can hardly hold it, wants to fall out of your hands because the sidewalls are so weak, right?

Uh, so it's just continual competition. I like to use the beverage container business as an example of each one doing more and more. The glass people are starting to run into some problems um being competitive, okay. The aluminum and plastic have gotten thinner and thinner. But if you go to Japan — haven't been a lot of you in Japan — at least 15 years ago they were still using steel beverage cans. Them in Europe 10 years ago, they still use them in Japan for their — they have hot drink machines. And the reason is the steel lobby in Japan. I mean aluminum is clearly cheaper than steel for a beverage container. But the Japanese have a big steel industry and they have no aluminum industry to speak of, and so it's basically again tariffs, restraint of trade. Um, you just can't get in to do business there.

In fact, um, it doesn't really matter now cause we're in digital photography, but when I was there, um, about 5 years before — around 1980, or the late 70s or something — Kodak um wanted to build a plant in Japan to make film, okay, photographic film. And um in order to build a plant in Japan you had to get permission from the Japanese government. And the Japanese government has never said yes or no to Kodak. Of course Kodak's about bankrupt if they're not bankrupt, okay. But um, but at the time it was estimated that if Kodak could have come in and started producing film in Japan, they could have kept Fujifilm from ever becoming a dominant world player. But the Japanese government just never answered Kodak's request of, can we build a plant? They're still waiting for an answer. Of course Kodak doesn't want to build a film plant anywhere, no one wants to build a film plant anywhere, but in fact, you know, from 40 years ago we're still waiting for a response, okay.

But that delay basically um allowed Fuji to get, take a world position that was able to compete with Kodak. And eventually Kodak lost its monopoly on, you know, making film in the world. They didn't have a monopoly — Ilford and others in Europe — but they had a near monopoly. And if you went to Kodak in the mid 1990s, uh Kodak, the people at Kodak had grown up in this artificial environment, a semi-monopolistic environment, and they didn't know how not to spend money.

I mean I had some students do thesis at Kodak, and I remember going to Kodak Park and taking a tour. And they had — Kodak Park was built by um, what's his name, founder, Kodak — Eastman, yeah, Building Six is the Eastman Building. He got MIT out of bankruptcy in 1917. Uh, anyway, the Eastman, George Eastman built Kodak Park as a manufacturing facility back, you know, in the 1910s and 1920s. And it wasn't laid out the way we would lay it out today, so they had a loading dock and stuff would come in, and they had to move boxes about a half floor up. And so instead of going out and buying some conveyor, you know, conveyor tables to move the boxes and then a little elevator to move them up to the next level, Kodak designed and built — out of stainless steel, not carbon steel — their own conveyors and elevators, because they were used to, you know, keeping full employment. They were basically printing money. They called it film but it was printing money. And they had no real competition. And the people at Kodak just burned through money, because if they showed that they were too profitable, the government'd get all upset and come in and break up the monopoly. So they just burned through money like it was water, okay. And then when they did get a competition, these Great American businessmen didn't know what to do, and so they kept burning through money.

I worked for Bethlehem Steel in an era like that, okay. Bethlehem Steel in 1973, the year before I went to work for them, most profitable year ever, okay. And the American steel industry after World War II — the American — U.S., the American steel industry owned 75% of the world steel making capacity in 1945, cause we had bombed out most of the rest. It's great, you just bomb out your competition. There's a good businessman, okay, when you can bomb out your competition you can take a world leadership role. Well, I went to work for them in '74, which is 30 years later approximately, and all the managers who had been hired as young employees in 1945 were now running the company. And they still had the attitude that they controlled 75% of the world steel business and that they could dictate prices and everything else. Well, when I went to work for them, we controlled 25% of the world steel industry, okay. But they still thought they controlled 75%, because that's what they grew up on, okay. And uh, it was decreasing and continuing to decrease.

And I went to um a training for the 500 new college employees of Bethlehem Steel in 1975 — let's see, it was uh summer of '75. We went to this two-week training in an auditorium, and we had countless vice presidents come in and give us talks. And this one Financial vice president comes in, he says — we are, they just had their most profitable year ever two years before, right, a year and a half before — he says, we [are] one of the most efficient steel companies in the world, cause our blast furnaces were built in 1912 and our coke ovens were built in 1911. And I, out of 500 idiot new newbies, I raised my hand and said, I don't understand why having old equipment makes you more efficient. He says, because it's fully depreciated. I said, oh. And I realized that, and he made some comment about I didn't understand finance. So a year later I took a finance course and I learned I didn't understand finance, but neither did he, okay.

Anyway, now you heard my tirade on American uh manufacturers and businessmen. They're not as great as they think they are, okay. Let's take a 5 minute break and we'll come back and [end].