by David L. Morton, Jr. , Sept. [date ?], 1993 via telephone
Note: Bohdan Kostyshyn (1927-2018) was a Ukrainian-American engineer who spent much of his career with IBM. https://www.arnmortuary.com/obituaries/bohdan-kostyshyn
Transcript
[This recording is incomplete. The recording begins mid-sentence]
Kostyshyn
Yeah, I mentioned that. I’ll talk about other memories in just a minute. The analog industry and the video industry had no such problems. And the video industry, yeah, and the analog industry both used… You can purchase plated tapes. And that basically is, I believe, and I’m not sure, I cannot be quoted on this because I really don’t know. But I would expect it’s some type of cobalt nickel phosphorus plated on a backing like mylar or something like that. But the analog industry doesn’t give much of a damn whether there’s a small void occasionally or not, because the information, you never see it in a picture or in music. And the advantage of this particular kind of tape, the metal tapes, is that you can get much higher coercivity. And because of the higher coercivity, you can record shorter wavelengths or shorter bits on a magnetic recording medium for a given velocity. And that’s what the entire thrust has been in the analog industry, to be able to record at basically higher, for higher, at higher and higher frequencies without increasing the velocity of the magnetic recording medium. Back when Ampex was in business, or they still be in business, they sold an audio recorder and they had these big 10 inch reels. I don’t know whether you’ve seen pictures. They’re kind of antiques right now. But you would, you would use the tape at two different speeds. And in fact, on your video recorder, you probably have an extended play on your video recorder where you can get eight hours on the video. Do you have a video recorder? You can get up to six or eight hours on the video recorder, but if you have it on the other setting, it will record only two hours. Well, all that means is that the tape is going three times as fast, and thereby allowing a given frequency to have a longer bit recorded on the medium, and you have a better signal. And in fact, you can record higher frequencies than in the extended range. That’s all that that is. So that’s what the entire thrust has been. Now, let me take one step further back to digital recording. As I said, up to about the 60s. or thereabouts. The trust has been to increase the bit density of the magnetic recording medium.
But then all of a sudden, people realize that that’s not the only dimension. There is another dimension. You have a linear, which is in the direction of the magnetic movement, or the movement of the tape, but you also have across the tape. Typically, the width of magnetic recording heads, it has a high track density would be 50 tracks to the inch. Well, I recall back at one of the magnetism conferences, Jeff Pate, who is a friend who used to work for IBM, and who had been, and Bates, by the way, is a leader in magnetic recording, or had been, in magnetic recording materials, basically. I think he works, if he’s still working, he’s working in Phoenix. for either Phoenix or Longmount, which is right near Boulder, Colorado. He recorded on the magnetic medium by recording and then moving his head over and moving it over and moving it over. And where he recorded a thousand texts of the entrance. And he called it the world’s only write-only memory because he couldn’t read it. He could look at it with what are called pitter patterns. Back then, what you had was a liquid that had very, very fine particles in it, and you could put it on your magnetized region. And you could look at even domain walls. So he could look at that, but he couldn’t read it. And what people started doing was looking at techniques of increasing the track density as opposed to the heat density. And they did this by, instead of fabricating magnetic heads, of permalloy by making a lamination similar to what you had in your transformer cord, and making these laminations up in other states. Well, making a ferrite head, and later on, more and more people with a ferrite head. They began evaporating heads, making heads with evaporated permalloy, so they could control the width of that head very accurately. Now, I got out of the business shortly after that particular point. The guy who, in IBM, who did a lot of work in that area, and in fact as an IBM fellow, and I don’t know if he’s still at IBM or not, but he was located at IBM Yorktown. His name was David Thompson. And I knew David quite well. David did work in, he was well recognized for his work in plated-head technology. By the way, there’s another person that was in IBM Yorktown who did a lot of work in magnetic materials. And his name was Lubyomir Romankiw. His R-O-M-A-N-K-I-W, which is Ukrainian, like I am.
But David did work in increasing the track densities. And if you look at the capabilities that you have now on on hard disks, on computers, you have on a five inch disk, or no, it’s smaller than that, three and a half inch disk, you can get 20 to 30 megabytes of information. That’s an enormous amount of information, and digital information nonetheless. So they have to be almost, if not in contact with the heads, and they have to have a very high truck density. But you can imagine the problems that you have on alignment of the heads with regard to a given track. So there have been a lot of, there’s been a lot of work on alignment of heads in a given track because on the disc systems you have one head and you must traverse the magnetic recording medium until you get to the track and then you rotate the medium. I’ve talked a long time and what questions do you have?
Morton
Actually, I was just, since I didn’t know much about . . . [gap in recording]
. . . sensitive to the fact that IBM is. . . the only reason I told you as much as I have, and I’ve tried to stay away as much as possible from products, and I would appreciate no reference to products, specific products. Although you can mention the 650, because that’s very early, very early work. But and you can mention other things that that are referred to in the literature, but please don’t mention the military products that I referred to because I don’t know how sensitive that is. And I don’t believe it is sensitive, but I also don’t believe that you should cite those particular products.
Morton
Okay. I understand.
As specific products. You can say in the military and that kind of thing, but don’t refer to a given specific product. Okay. And you can even refer to the problems with regard to stress, vibration, et cetera, et cetera, et cetera. By the way, they had magnetic recording in all kinds of situations, and they had all kinds of magnetic recording, tape, drum, and disc. The largest systems you could imagine, could go on a naval ship of some kind. That sort of thing.
And I’m sure that they still have that. As far as I’m concerned, my career is simply this. I started off working with the Bureau of Standards. And I left the Bureau of Standards after a couple of years, and I went to work for IBM. And I worked for IBM for about 35 years, 34 years. And most of the, about half of that time was as I liked to put as a worker, and the other half of the time was a manager. When I was working at [unintelligble], that’s when I was involved in my medical courts and such. About three quarters of the way through my my career, I decided to look at other kinds of technologies, and I got involved in other types of technology, inkjet printing, things like that. The majority of the work that I did, and I published a couple of papers on magnetic recording technology, or magnetic recording theory, not technology. I was mostly involved in magnetic recording theory, and then my role was to help design magnetic recording systems. And for that, as a result of the theoretical work, we were able to make better use of the material, materials that were made available, materials with higher coercivities, materials with, you know, ferrite-head materials, things like that. And so that’s really what my role was. The school that I went to, I I went to Queens College in New York for a bachelor’s in physics and to Syracuse University for a master’s in physics. And that’s as far as I went.
Morton
This may sound a little prosaic, but could you tell me how to spell your name? I never did get that.
No, no, no, that’s no problem. It’s not easy to do. K-O-S as in Sam, C as in Tom, Y-S-H-Y-N. first name is B-O-H-D-A-N.
By the way, if you have questions, you can look at magnetic recording as basically a three-step process. And the first step is where you have the magnetic head write head, W-R-I-T-E, in contact or close contact with the magnetic recording medium. And you You do the same thing with the right head that you do with a transformer. You put current through a coil. And this in turn, at the gap of the right head, causes what is known as fringing fields. And these fields in turn align up the magnetic moments in the magnetic recording material as the material goes by. Then what you do is you reverse the direction of the current. In the analog recording, What you do is you, it’s not an on or off kind of thing, you know, it’s not one direction or the other, but it’s different levels of current. It’s not quite the same, but it’s similar.
The second thing that happens is that the area that you just magnetized leaves the region of the recording head. Now, I’m sending the paper to you that notes. . . you can read through it and you may want to and you may not. But what happens is you get poles generated at the regions where you reverse. So if you get a bit, all that a bit is in digital recording is an area that’s magnetized in one direction. Then you magnetize the next area in the opposite direction. Got that? Do you have that?
Morton:
Yeah.
Okay. Now, at the intersection, the transition, you have what are called poles. And these poles, in turn, it’s like having a bar magnet. You have an external magnetic field. And that’s what you sense when you read again. But there’s also an internal magnetic field that’s opposed to the direction of magnetization. In other words, if you look at the arrow pointing in one direction and look at the head of the arrow as being the north pole and the south pole on the other end, the north pole generates a field that is in the opposite direction. So it tends to want to demagnetize or to drive the moments in the opposite direction. Now, if the material has a high enough coercivity. .. rRemember, it’s the width of the hysteresis loop. The field will be there, but the magnetization will basically remain in the same direction. If the coercivity is low, the magnetization will be reduced a great deal. Then you’ll basically demagnetize that bit. There still will be some remanant magnetization pointing in the same direction, but it will be very low. So what you want to do is to have a material that has a high enough coercivity. And by the way, the number of poles that you generate is going to be proportional to the thickness of the magnetic material. [end of tape]
Side 2
Kostyshyn
OK. All right, let me just get comfortable here because I’m at my desk. OK. You know, I’ve been thinking a little bit about what you set out before you. And you’re a history major, right?
Morton
Yes.
Kostyshyn
And what you’re looking at is the history of magnetic recording or what? Or magnetic recording media.
Morton
Well, magnetic recording generally. I’m interested in the commercialization of it, the building of the magnetic recording industry.
Kostyshyn
OK. Yeah. Well, in order to– there are several things here that you really have to consider. You’re covering a very broad range.
Morton
Yes.
Kostyshyn
For instance, you can separate magnetic recording into several major categories. One would be digital recording, which is what is used on computers and other machines like that. And in fact, some of the, one is on analog recording, and that’s the type of recording that had been that you record music and that sort of thing on. And one is video recording. And on video recording, of course, you record pictures and other type of data. And the requirements are generally different in each category, as you would expect. For instance, in digital recording, generally, you cannot tolerate what are known as voids, because a void would give you a false reading. Now, the digital industry gets around this somewhat by having what is known as a parity bit, and they can determine whether or not they have problems in a piece of information. in what is called a bite. But in general, the requirements for digital recording is that it be error-free. Whereas in analog recording, and even digital recording, today they have digital recording of music, as an example, digital digital taping, and that sort of thing. But even there, you don’t have quite the problem that you do with digital recording for digital information. By digital information, I’m talking about bits in computers, for example. Because you can drop a bit of information in an analog signal, like music, and you’d never know it, even a video signal. You just never know it. Whereas you cannot drop a piece of information in in digital. When I speak of digital, I mean computer-type data. So the requirements are quite different. But in general, both techniques, all of the technologies have tried to do one thing, and that is to increase what I would call the density of recordings.
Back when magnetic recording first began, actually magnetic recording began with it, and I’m just relying on memory here because I really didn’t look this up, but it was a Dane, and I forget the name of the Dane who really invented magnetic recordings. Poulson was his name. Yeah, that sounds correct. I recall, now I’m 66, I recall when I was in the Navy in 19, oh, about ’48 or thereabouts, I came across, I happened to be in a Naval school at the time, and they had a wire recorder, and back then they were recording on wire, and I think Brush, in its early days, recorded on wire. By the way, I had some communication with Brush in my early years. Do you? Yes. In fact, I had them make up for me a set of special recording heads, because at that time I was deep into magnetic recording theory, development of magnetic recording theory. And I did get a series of magnetic recording heads from them, but not much ever came out of that. Basically what I had them do was to make a series of heads that had different pole gaps on them.
Morton
Yeah. Since that came up, I was wondering if I’m aware of some research that they did in about 1945-46 on some plated I want to say they were disks. They might have been drums, but they were for data storage for computers. And it was some, you know, military contract work. And I think the reason I brought that up the other day was I was wondering if there was a connection between that research and the stuff that went on.
Kostyshyn
No, no, no. Not at all, because what happened was that at that time, all of the companies, and IBM included, were looking for ways to record data. digital data. Now, by the way, I spent approximately half of my IBM courier working in the, what you would call the military end. I started in commercial, then transferred into the, you could call it military, but it’s actually government, government-based. The difference being that in in that branch, the work that was done was done in response to, in response to what are called requests for quotes, RFQs, which really is a request for work to be done by industry, by the government. And there are several things in the military that The people on the commercial did not… There are several requirements that the military had that the people in commercial did not have. And I’ll tell you. One is the thermal environment or the environmental conditions that you had to operate under. And usually the thermal environment was rather stringent. It was usually from about minus 40 to plus 75 and even higher centigrade or Celsius. And that’s a very stringent environment. And the other stringency or the other requirement is a vibration and a vibration environment and a moisture environment, all of those things, because most of the military equipment had to operate for instance, on aircraft or even in space. And there is a third requirement that may not occur to you. And that requirement is, and I mentioned it the last time, that you have two things. One is a memory that will not lose its information when the power is shut down. And the second one that the military wanted was a thing called non-destructive readout, which means that you read the information and do not destroy the information that you’re reading. And most of the technologies that were available then had a destructive readout. There was a technology that occurred in about 1960 that you may be familiar with or not familiar with, and it was called bubble memories. Are you familiar with that?
Morton
Yes.
Kostyshyn
Okay, bubble memories, and I worked on that some, too. Bubble memories essentially were magnetic memories that used a special material. It was called a garnet. It’s a ferrimagnetic garnet, and they were built They were garnets that were doped with, by doping, you put in atoms of material in the structure. They were doped with rare earths, like yttrium, one of the materials was yttrium iron garnet. The technology was first developed by Andrew Bobak in Bell Labs. And the industry got very excited about this. And a lot of money was invested in in this as a substitute for magnetic recording. Now let me just tell you, give you some picture. Do you know what it is?
Morton
The bubble memory system? I know sort of generally of it, but…
Kostyshyn
Well, all that is a magnetic recording medium where the easy axis of magnetization, in memories, you have what is known as an easy axis and a hard axis of magnetization. where the bubbles– if you look at the material, and you could actually see the areas because you could look at the materials using either the Faraday effect or the current magneto-optic effect. And essentially, the materials were prepared in a furnace. at very high temperatures, and it was required that the materials be single crystals. And what happened then was you sliced the material like you would slice a piece of paper. And you had like a wafer. It’s something like the technology that you have for semiconductors. It would be the same thing. You have a wafer like you have for semiconductors. But what you had to do was to place a pattern on the surface of this material. This material had to be single crystal and it had to be without defects. Now what you could do is to form what they call bubbles, but the reason they look like, the reason they call them bubbles is that they looked round when you looked at them. And you could look at them with the. . . Generally, with this material, it was with the Faraday effect. These materials were transparent in the infrared and in the visible red range. And if you shone light through the material, if you had areas of different magnetization, in other words, if you can envision a magnet going north-south, vertical in a sea of north-south, vertical, and the opposite direction, then the light, the light would have to be polarized light, and the plane of polarization was rotated differently for the different areas, and you could analyze and actually see these areas. But the point is that each of those bubbles, each of those round areas formed a bit of information, and you could move the bubbles around, Effectively, on a magnetic recording medium, what you do is you magnetize bits on the tape, okay? And you move the tape. In a bubble memory, you magnetize bits on the bubble medium, and you move the bubbles. You kept the medium still.
Morton
Oh, yeah, I see.
Kostyshyn
Okay, now the advantage of this, and the military people were very excited about this, was that it was a non-mechanical memory. The tape recorders in satellites were notoriously bad with regard to breakdown. And it was the first thing, basically, the meantime before failure, MTBF, it’s referred to, is high, is low for the tape recorders. To my mind, to my way of thinking, one of the most outstanding and I mean outstanding accomplishments of the ’50s, late ’50s and ’60s was the sending of a satellite to Mars. Not Mars. Was it Mars? Yes, Mars. It was to Mars taking pictures, recording those pictures on magnetic media. and then transmitting that data back in a very slow mode. And you can look this up in the Scientific American and other. It was just an incredible, an incredibly complex thing to do. The tape recorders worked, the antenna worked, the whole business worked. And I cite that as an example of magnetic recording that was Just, in my mind, superb. Okay, but getting back to the, getting back to the thrust of magnetic recording, back in about the 50s, which is really when I came into BDC, the thrust was to increase the density of information that you could store on magnetic recording. And I have to concentrate on digital because that’s what I know most about. But about the same time, I think Philips of Holland came out that the idea of the cartridge, you know, the cartridge that you use today, that’s a Philips invention. You know what I’m referring to.
Morton
The cassette.
Kostyshyn
Cassette, right. Now, also about that time, there was a company, I don’t know what the status of the company is today, I don’t hear too much about it, called Ampex, who began recording video information on magnetic tape. I’m just looking at something here, just one moment. Yeah. If you write this equation down, the frequency is equal to the velocity divided by wavelength. Okay? That’s…
Morton
I don’t know all the right symbols for those, but…
Kostyshyn
Well, F is equal to… I’m looking to see whether I have the equation correct. I’m looking at the… I’m sorry to keep you waiting.
Morton
That’s all right.
Kostyshyn
Yeah. Frequency is velocity divided by wavelength, right? F is equal to V divided by lambda. Lambda is the Greek letter that’s generally used for wavelengths, okay? Now, the problem that they have with regard to video recorders is that the wavelengths, the frequencies are very, very, very high. So essentially what you had to do, in order to record reasonable wavelengths, which you could equate to bits in length, you know, an actual length, on a magnetic recording medium, you had to increase the velocity to very, very high values. Do you have that equation that F is equal to V divided by lambda? Mm-hmm. Okay. So you had to increase the velocity at the very high values. And Ampex came up with a very neat scheme where they moved the tape and they had a rotating head that rotated at an angle relative to the motion of the tape. where one scan on the video, you know how a video system works, you have a series of scans, where one scan was equal to one scan across the tube, one scan across the tape. But the heads were moving at very high velocity. The tapes were about, oh, I think about an inch, an inch and a half, maybe even two inches wide. And essentially, the video systems that you have today, the VCRs do the same sort of thing, I believe. Are you there? Yeah. Okay. So Ampex was a leader early on in recording very high frequencies in an analog system, because video recording and analog and music is basically the same thing, except that the frequency The frequencies for video recording are very high. The frequencies for audio music are much, much, much lower. But right around that time, IBM really got into the magnetic recording. aspect of it, and they started with their 650 headers, a 650 system, as I told you the other day. They had other drum systems, and I believe those other drum systems were developed in Rochester, Minnesota, and they came after the 650. I was involved with drum systems when I was in the military, and in particular, we designed two systems. One was, as I mentioned, the Titan recording system, which was a relatively small drum, And you have to understand, when I told you the environmental aspects, this thing had to be fired up in the rocket. So you could see that it had vibration, it had thermal, all kinds of requirements on it. So effectively, it was a relatively low bit density or If you want to refer to frequency, it was a low frequency system, but you kept it low, it was kept low so that it would withstand the environment that it was placed under. Now, let me point out one thing. One of the most critical, I’m looking for the word, parameters, One of the most critical considerations in magnetic recording, any magnetic recording, is how close you can come to the magnetic recording medium with your magnetic recording head, particularly your reed head, but your write and reed head. And in tape systems, you literally are in contact with the magnetic recording medium, with the head. But in gun and disk systems, that is not possible. So there had been a lot of, there had been a lot of work, there had been a lot of work on developing what are called flying heads. And in fact, most of the heads that are in computers are literally flying heads. And they’re flying over the magnetic recording medium at a height of, say, a millionth of an inch, very, very, very small distances. The larger the distance, the more problems you have with regard to with regard to how short a bit or how high a frequency you can write on the recording medium. So, you know, the magnetic recording medium, development of magnetic recording medium is not independent of the development of magnetic recording technology. And the thrust in the ’50s and going on into the ’60s, I would say, have been to develop the capability of recording higher and higher densities. In other words, the linear densities. The bits became shorter and shorter and shorter and shorter, to the point where we designed a system that had 2,000 bits per inch. There were tape systems and disk systems developed in San Jose. There was a large group of people in IBM developing Magnetic recording in San Jose, and they were developing recording systems that had 7000 bits, but of course… The military requirements, again, and the commercial requirements are quite different. And when you put a different set of boundary conditions on such things, such as vibration, et cetera, et cetera, you automatically make things more difficult as far as how close you can come to the magnetic recording medium, and then that, in turn, determines how high of a density you can About after the ’60s, I think. About after the ’60s. I’m not sure. I’m not sure of this information. By the way, there was some work being done at IBM Yorktown. And I knew most of the people in IBM who was working on magnetic recording, and they in turn, I think, at least some of them knew me. We communicated quite a bit. But let me get back to the other system that But we developed, we developed a system called for a naval aircraft called S3A. And S3A is a naval aircraft that is a submarine detection type of aircraft. And they needed a 50 megabit, 50 million bit drum. And we had developed that in IBM Owego. Owego is where the military, IBM had military plants at Owego in the Washington area and also one on the west coast. And that used, as I told you yesterday or the other day, that used the material that was basically cobalt nickel phosphorus. And it had a thickness of magnetic recording media of five micro inches, that’s five millionths of an inch. And on top of that, you had a protective layer of gold and rhodium. The rhodium was placed on there to give it a hard surface because the heads were in contact when the drums started up. But shortly after the drums started up, the heads flew. They were designed to fly. Otherwise, it was like a big brake drum with a head per track. And the gold was placed on there to protect the magnetic recording medium from the rhodium bath. The rhodium bath was very acid, so you couldn’t plate without destroying the magnetic recording. That was the system that, that was the last system that I know of that was developed in IBM Orego. And after that, I think the military began looking more closely at using semiconductor memories as opposed to other types of memories. They spent an enormous amount of money, I’m talking about the military now, with various companies, looking at ways to make the semiconductor memories what is known as non-volatile. In other words, that’s the term that it’s used for. When you lose power, you don’t lose the memory. But I’m not quite sure at this point as to what is being used because I’ve been out of it for quite a while now. Let me get to the magnetic materials for a moment. Okay. In magnetic materials, I have some notes. I remember, do you recall I told you I gave a talk on magnetic recording? years ago at SUNY Binghamton. Right. I have those notes and if you want, I’ll copy them.
Morton
Yeah.
Kostyshyn
They were all handwritten, but they should be pretty clear.
Morton
Yeah, I’d be very interested.
Kostyshyn
What is your [address]?
Morton
The address is 1598…
Kostyshyn
You’re up? About two weeks. of magnetic materials, the requirements for magnetic materials. There is a theory that was developed by two people, Stoner, S-T-O-N-E-R, and Wolfarth, W-O-H-L-F-A-R-T-H.
And are you familiar with magnetics at all?
Morton
Well, some. I mean, I’ve–
Kostyshyn
Are you familiar with domain theory.
Morton
Yes.
Kostyshyn
Okay.
Kostyshyn
Well, you know that in magnetic materials, you have small regions of magnetization where, basically, you have small regions where the magnetic moments, magnetic moment is like a small magnet. where the magnetic moments associated with the atomic, the atomic magnetic moments are lined up all in one direction. And that region is called the domain. Now, when you magnetize a piece of material, essentially what you do is you line up larger regions of the magnetic material so that you have all of the atomic moments lined up in one direction in a given area, okay, so that you have poles defined north and so forth.
Kostyshyn
Now, what [Stoner and Wolfarth?] did was to examine the requirements based on the substructure of magnetic materials and Well, it’s really more of what is called a macro structure. And they found that you can get movement of, you can get magnetization by one of two methods. You can get magnetization by so-called movement of domain walls. And the other thing that you could get was the magnetization by rotation of the magnetic domains within the structure. Now, they pointed out that when you get down to a certain size, very small size, physical size, domain walls are not supported so that you have the magnetization pointed one direction or another. And if you can imagine the oxide, the ferric oxide, these are very, very small particles. So they tend to become single domain particles or at most one or two domains in a given particle. And the way that these magnetize is typically by rotation. And the shape of the hysteresis loop that you get from this is sort of square. In fact, in an ideal particle, it’s perfectly square in the direction of the easy axis of magnetization, which is the axis that the magnet point . If you go in the other axis, you’ll find that the hysteresis loop looks like a straight line, actually. I have some of that in the notes that I’m going to send you. Now, the whole thrust of the development of magnetic materials has been to develop materials that require higher and higher magnetic fields in order to cause rotation. That’s called the coercivity. Now, when you talk about individual domains, you really can’t talk about the histories. You can only talk about a hysteresis loop for an average of, it’s an averaging process. It’s an average way of looking at the magnetization of the material. Now, the entire thrust has been to develop these materials, and for the longest time, I believe it’s ferric oxide, that’s it. There’s two oxides of iron, ferrous and ferric, and I believe it’s ferric oxide, but it’s the magnetic material. And the first magnetic media, believe it or not, was made by paint companies because they had, for pigments and paints, they were using ferric oxide. And one of the things they were trying to do was to create uniform surfaces. You can imagine the problem that you have trying to get a uniform thickness of material by using pink type materials. And it’s a real dog of a problem, but that’s what was used for the… And in fact, that’s what’s used on magnetic recording tape today. It’s a coating of ferrous oxide. The researchers tried to look into ways of making the materials have a higher coercivity. Now, again, remember, the coercivity is the width of the history of slope and the remnants or retentivity or whatever, saturation magnetization is the height of the history of slope. They were looking for a material that had a high coercivity and that was relatively square in shape. Now, some of the things they did was to dope the oxides with other materials. And I believe, and I can’t be quoted on this, you have to look into the books, they use manganese, materials like manganese and other such materials. And they were able to, in fact, there was a material, I think it was manganese dioxide, that had no, that was ferromagnetic. But it had what was known as a very low Curie point. Now, are you familiar with the Curie point?
Morton
Yes.
Kostyshyn
Yeah. Okay, the Curie point is this. If you heat the magnetic material up, it basically lose the thermal, the energy, remember, in order for the, atomic moments to be lined up on an atomic scale, there’s a certain amount of coupling that takes place between that. There’s a certain amount of energy involved in that. Do you follow that? So that there’s this coupling that takes place between neighboring atoms. Now, if you take any magnetic material, any magnetic material, iron, nickel, cobalt, and the compounds, and heat it, You’ll get to a point where the thermal energy involved, because what happens when you heat it, the atoms vibrate in their lattices. And the coupling, the thermal energy breaks down the coupling between the atoms, and so you lose the magnetism. In fact, you can demagnetize a piece of material by heating it. If you have a screwdriver that’s magnetized as an example. All you have to do is take a torch to it and heat it and you will demagnetize it. It will become magnetic if you stroke it with a magnet so that you will line up those domains again. Now, let me take another step back because this has to do with material theory. There are various classes of magnetism. Are you familiar with that?
Morton
I don’t think so.
Kostyshyn
Okay. There is a class called ferromagnetic materials.
Morton
Right.
Kostyshyn
And these generally are materials like iron, nickel, cobalt. And these materials on an atomic level, all of the magnetic moments. Now, you understand what I mean by a magnetic moment? Okay, you could look at it as being a little arrow associated, which indicates the direction of the magnetization. All of the magnetic moments point in one direction on the axis. There is a group of material called ferromagnetic materials, that’s the RRI. And these materials have magnetic moments pointing in both directions. the majority or a large portion pointing in one direction. Now, the saturation magnetization, which indicates the height of the hysteresis group, is higher for the ferromagnetic materials than for the ferromagnetic materials, simply because some of the magnetic moments are being canceled in the ferromagnetic materials. There is a class of materials called anti-ferromagnetic materials, where There are as many magnetic moments pointed in one direction as there are in another. So you have a net value of zero. But there is a coupling nonetheless. And then there’s finally a diamagnetic material. And then that’s not important for your point of view. But the ferromagnetic materials have become very important because they are basically oxides. And many of the magnetic recording heads that are made today are made from ferrite magnetic materials. And they have certain definite advantages. Because they’re oxides or because they’re crystalline, they are non-conductors. And one of the problems that you run into when you, have you ever taken a transformer apart.
Morton
Yes.
Kostyshyn
Okay, the transformer is made of leaves.
Morton
Right.
Kostyshyn
The core, right? The reason they do that is because when you send current through the primary coil, what essentially you’re doing is you’re lining up the magnetic moments in the transformer, and this causes a change in the magnetic moments under the secondary coil, and by Faraday’s law, you get a voltage And you have to use these with alternating currents. So you’re energizing one direction, then the other, and then back and forth and back and forth. Well, one of the things that happens is because the core material itself is a conductor, you have generated in the core things called eddy currents. And these eddy currents create a back electromotive force. And they create heat. And the eddy currents are in kind of the vertical direction. So the way you get away from that is to shorten the pad blast. And when you do that is you make the transformer out of leaves and insulate between the leaves. And thereby you cut down on what is called eddy current loss. But if you have a transformer that has a ceramic– a non-conductor core, then you don’t have any cards. And with the ferromagnetic materials, these are basically ceramics. And they are non-conductors or very, very low conductors. So you can make magnetic heads without having to leave them. Okay, you understand that?
Morton
Yeah.
Kostyshyn
Okay, so that’s the different classes of materials. Now when you get to platings. Platings work because essentially what you do when you plate is you create islands of magnetic material separated by barriers of non-magnetic material. For instance, and I’ll just refer to the one system that we used in the and that was cobalt nickel phosphorus, pretty well-known bath, pretty well-known magnetic material. But by varying the current density and varying the composition of the bath, you can create islands of cobalt and nickel separated by cobalt phosphide and nickel phosphide. Now the cobalt phosphide and nickel phosphide are non-magnetic. So that you effectively have, like you do have on the ferric oxide, you have small particles. And it’s important that you’re able to control the size of those particles. And that’s how you get the high coercivity. Now the problem with platings is that you’re very subject to the surface that you’re plating on, and you can get voids. Now, the void, if the void is small relative to the bit size, you understand what I mean by bit size? Okay. Then it doesn’t matter too much. But if the void is of the same order of magnitude as the bit size, you’re in deep trouble because you have a, you have a, you have a defect. And that cannot be used for for digital recording, because you run into defects and then you run into errors. So the, in general, IBM looked at, and I recall this, and we did this as well, in a week ago, we looked at plating, plated tape. In order to plate tape, you have to have first the medium, a backing. And then you have to have a way to plate on that. Well, the first tapes that were used were sort of like a metal medium, like bronze or flexible copper or something like that. But you can make that only so thin. It becomes bulky and it’s not flexible and very difficult to deal with. So the next thing that was looked into was plating on Mylar or other such materials. But in order to do that, it’s a very tricky thing. It’s not a good surface to plate on. Because the way you do it, you sensitize the plating surface with, and I forget, palladium, palladium oxide, palladium, some type of palladium bath. I don’t know the bath. First you do that, then you have to plate on the palladium. The palladium fastens itself to the Mylar. And then you plate on that with a thin layer of copper, and then you plate on the copper surface, then you can get a plated tape. But you’re inevitably, you run into the problem of voids. And try as you may, it becomes very, very, very difficult to create error-free surfaces, which, by the way, was the killer for, there were two killers for bubble memories, and I mentioned that, and I’ll talk about bubble memories in just a minute.
[recording ends]