Difference between revisions of "Oral-History:Ron Gedney"
|Line 271:||Line 271:|
<p>Well, that’s been very, very helpful to me, and it is really excellent background. Thank you very much for all your help. </p>
<p>Well, that’s been very, very helpful to me, and it is really excellent background. Thank you very much for all your help. </p>
[[Category:|Gedney]] [[Category:Engineers|Gedney]] [[Category:Corporations|Gedney]] [[Category:IEEE|Gedney]] [[Category:|Gedney]] [[Category:Memory|Gedney]] [[Category:Components, , & |]] [[Category:Power, & |]] [[Category:|Gedney]] [[Category:News|Gedney]]
Revision as of 17:19, 30 March 2012
About Ron Gedney
In addition to his significant contributions in the field of computer packaging and manufacturing, Ron Gedney served in a variety of positions as a distinguished member of IEEE CPMT society. As an employee of IBM, Gedney was involved in the procurement of components, the development of reliability standard tests, the production of chip and circuit technology, and the management of a thin film substrate program. He served as program chairman of ECTC and on the board of governors of CPMT. He subsequently ran and served as vice president and president of the CPMT. During his tenure, he helped to develop an efficient organizational structure and a successful packaging technical committee, as well as expand the CPMT awards plan and help restore CPMT’s financial security.
In this interview, Gedney describes shifts in computer manufacturing and marketing from the 1950s through the 1970s. He details his contributions to the IEEE CPMT Society. The interview concludes with assessment of Gedney's educational preparation for his career in packaging.
About the Interview
RON GEDNEY: An Interview Conducted by Robert Colburn, IEEE History Center, 2 June 1999
Interview #277 for the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc.
This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.
Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, 39 Union Street, New Brunswick, NJ 08901-8538 USA. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.
It is recommended that this oral history be cited as follows:
Ron Gedney, an oral history conducted in 1999 by Robert Colburn, IEEE History Center, New Brunswick, NJ, USA.
Interviewee: Ron Gedney
Interviewer: Robert Colburn
Date: 2 June 1999
Place: ECTC Conference in San Diego, CA
IBM, IEEE, and the computer industry, 1950s-1960s
Going back to when you joined the IEEE, what was the industry like at the time?
Well, I got my BSEE in 1957, and went to work almost immediately for IBM. I joined the IEEE on July 1st, 1957, which was about a week or two after I started with IBM. In 1958, IBM made a decision to switch from vacuum tubes to transistors for future computers. They sent all IBM engineers through a series of four courses on semiconductor physics, modern physics, wave equations and transistor circuits.
Who taught those courses?
<flashmp3>277 - gedney - clip 1.mp3</flashmp3>
They were taught by professors from Brooklyn Polytech, Syracuse University and RPI. I was working in Poughkeepsie, New York, and they would bring in whoever could come down to the Poughkeepsie and spend the day teaching a class. The universities at that point were ahead of us in terms of transistor theory. In the 1950’s, mid-size and large Computers used vacuum tube circuits. Smaller computers used relays. In 1959, we started seriously designing everything with transistors. The first big system that went out with all transistors was called the Stretch System internally at IBM. I don’t know what the commercial designation became, but it was built for the government, and was a “supercomputer”. Today it probably wouldn’t hold a candle to a PC, but in those days it was considered very powerful. From the Stretch system we derived the 7090 which used the same transistor circuits and very similar circuit boards. That was the first large IBM commercial computer to use transistor circuits and was released in 1960-61.
IBM Components Division
To this point, IBM had purchased all their componentry. However, in mid-1961, IBM formed the Components Division to develop transistors and packaging. The kinds of componentry envisioned for the next computer systems were not readily available commercially. In 1964, IBM announced “System 360” – a revolutionary family of computers using hybrid circuitry. I believe it was in 1965 that Fortune Magazine published an article about S/360 which they called “The $5 Billion Gamble”. IBM had literally “bet the company” on the success of this system. System 360 not only implemented a new computer architecture, but a totally new component technology as well.
The first products developed and subsequently manufactured in the new Components Division were hybrid circuits, multi-chip modules—four transistors or four dual diodes to a little half-inch alumina substrate with screened and fired resistors. The passive components on these substrates were only resistors to support the transistor-resistor logic that was used. I was involved in procuring the components from outside the company at that time. We developed, with the help of suppliers, the R-pac, the molded tantalum capacitor, small ceramic chip capacitors, and a variety of things that would go with the new circuit board. To support these pin-in-hole components with “high-density” pins on 0.100 inch centers, glass epoxy circuit boards were introduced.
The circuit boards were developed by another group in IBM’s Endicott, N.Y. facility. At that time, these dense boards could not be purchased in the production volumes we were going to need. We thought the production volumes were going to be huge, based on our 5-year forecast for the System 360, and we thought we had planned everything pretty well. The day the System 360 was announced, we sold the entire 5-year forecast. That was when we realized we were in trouble. We could not buy or make enough components to build the systems in anywhere near the quantities that our customers were asking.
Underestimation of computer market
So this was another case of the computer market being severely understated.
Do you remember by how much the amount exceeded your expectations in terms of a factor?
It was many times. All I remember is that the whole forecast sold the very first day we announced. I had planned in 1964 that I was going to need about 10 million R-pacs in 1965. We needed 50 million. I was going to need about 7 or 8 million tantalum capacitors; we needed 30 million. And so I spent that year expediting components from suppliers, bringing in new suppliers, qualifying new suppliers, and anything and everything we could do to get the parts in so we could get those systems shipped out. And of course the systems build just went up from there. Back in those days we all thought that there was a finite number of computers that the industry could absorb. What we didn’t really appreciate was that once you had the computer you found many new applications and uses for it, so that the volumes kept expanding and expanding.
Where did some of those components come from? Who were some of the out source suppliers, and what kind of scrambling did they have to do?
Several suppliers had to build their own new plants. The R-pac was originally sold to us by Sprague Electric Corporation, and then we brought in International Resistor Corporation (IRC), Erie Technological Products (now Murata-Erie), and CTS Corporation, which still makes R-pacs today, were a big supplier. Kemet, Union Carbide and Sprague made tantalum capacitors. Erie and Aerovox made multi-layer ceramic capacitors. There were other suppliers as well, but these were the main ones I dealt with. CTS had a very unique design for producing R-Pacs, which I believe they are still using. GE was building mylar and big paper capacitors for power supplies; and Sprague was also doing aluminum electrolytics as I remember, and ceramic capacitors. Those are just a few of the companies we were working with at that time. We had to ramp up, and of course we put tremendous pressure on them to build production capability. It was not simple. Those technologies were very complex for that time.
Although these components were designed specifically for IBM there were no restrictions placed on the sale and, in fact, we encouraged the suppliers to make industry standards around them and build a larger customer base – which they did.
Reliability, testing, and standards
What sort of methods were used for ensuring reliability and testing and making sure they were up to IBM’s standards?
That’s a good question, because back in those days there were no standards and no industry wide reliability specifications, other than military. And the military specifications were all based on hermetically sealed parts, so they didn’t do moisture testing or migration testing. We were working with unsealed or plastic encapsulated parts. So we devised a number of tests. The one that we devised at that time was a 90% relative humidity high-temperature, 70˚C, with bias. Later we used 85oC and lowered the humidity to 85% RH – which was easier to control and we did not get condensation in the test chamber. Of course, later on the 85/85 test became an industry standard. We started using humidity testing back in 1958-60 basically on resistor technology. We were qualifying carbon film resistors, metal film resistors, tin oxide, carbon composition, and so we applied the same tests to the new components.
Up until that point in time - in the early sixties, all the components we used were axial leaded components. They were all cyclindrical, with leads coming out of the two ends. We tested thousands of parts, and again they were not hermetically sealed; they were plastic encapsulated. So we went through the whole series of temperature humidity and thermal shock testing to determine their reliability. We were writing the book on reliability in those days. There were no standard tests, and we had to devise them.
Was that a scary feeling?
Well, the chief technical officer or chief engineer of one of our suppliers was in our lab one day and I showed him some of the experiments we were running. One of the things that we tried was to take some of their resistors from a dry ice alcohol chamber—you know, we just had a little beaker—and put it directly into a beaker of boiling water so it had about a 200˚ shock. These components would actually physically vibrate because of the thermal shock that they were undergoing, and you could see them just vibrating like mad in the boiling water. And he looked at that, and he looked at me, and he walked over to one of our technicians benches and he came back with a ball-peen hammer, and he said, “Why don’t you try this?” He was drawing the analogy of overdoing it to this component. We weren’t actually asking him to pass that. We were experimenting with what tests would separate the men from the boys. And some of the parts, when you did that test, the shell or case of the part would just fly apart and completely disintegrate because of the thermal shock we were putting it through. Actually his parts were pretty good. They more or less passed the test. But I thought it was an apt analogy to what we were doing to the components.
Extreme destructive testing.
Screened paste resistor technology, substrate technology
What were some of the unexpected glitches along the way, in addition to the supply numbers problem? What were some of the things that would come up?
A couple of the interesting things actually were good things. We had purchased a lot of carbon film and metal film resistors in axial leaded format, and they were terribly moisture-sensitive components. If any moisture got in there and formed a bridge path, under a microscope you could watch the metal or the carbon film disappear, and they would just catastrophically fail. Tantalum capacitors were hermetically sealed, and if there was a short the case would rupture and spew solder. Well, when we started evaluating the screened paste resistor technology, it was extremely robust. It was almost impossible to cause those things to fail Catastrophically. You’d have to hit them with a ball-peen hammer. You could scratch them, you could put moisture on them, and almost anything you wanted to do to them, but they were pretty inert. The resistor element sintered at about 850 C and had a glassy phase which rose to the surface and protected the metallurgy. They were very, very good components.
And this came in about when?
<flashmp3>277 - gedney - clip 2.mp3</flashmp3>
This came in the early sixties. Dupont was the original supplier of both resistor pastes and electrode pastes and we did a lot of work with them honing and qualifying the technology. As time went on, the pastes shifted from the precious metals (gold and platinum) to the less expensive palladium, silver and ruthenium. At first we used Dupont inks, but because they were using a lot of precious metals that were relatively expensive we put a group together and started developing our own Later we licensed Dupont to manufacture the pastes for us. The transistors and dual diodes mounted on these substrates were 0.015 inch square, and had three little 0.005 inch copper balls, that we reflow soldered to lands on the substrate. So we were actually doing flip chip, but with three points and copper balls back in 1964. The substrates were approximately a half inch square, and we could put four transistor or diode chips with a the necessary resistors to make up what in those days was a whole circuit.
Controlled Collapse Chip Connection (C-4) technology
When we started to design with integrated circuit chips, along about 1965-66, we suddenly needed 16 balls or interconnections. However, we found that any non-panarity of the device meant that one or more pads did not make contact with the substrate and subsequently a good electrical connection. So one of our engineers, Lou Miller, invented the C4 joint or Controlled Collapse Chip Connection (four C’s), which later became known to the industry as flip chip technology. Basically, C-r replaced the copper ball with a high temperature solder ball on the chip. When raised to appropriate temperatures, the solder ball melted the chip settled into place and all the solder connections were made simultaneously.
The semiconductor device was placed on a viscous flux on the substrate over the top of the electrodes they were to join. We could watch the reflow process under a microscope during reflow—which we did it through a glass slide—and as the solder balls melted, and the surface tension of the solder pulled the chip right down about 0.001-2 inch. The solder balls or pads were only 0.005 inch in diameter, so this is quite a bit of movement. You could see the chip settle and then reach equilibrium and come back up a little bit, and then everything would stabilize. At the same time, the solder surface tension would pull the chip into alignment if it was twisted. We found experimentally that a chip would self-align if the pads were placed within ½ their diameter of the matching pads on the substrates. So, if you were using a 0.005 inch pad, you could misalign the chip up to 0.0025 inch and still get good interconnections. This was a tremendous invention, because now all of a sudden we had a process that was very, very forgiving in manufacturing. The technology has been greatly extended over the years. When I retired in ‘92, we were experimenting with 3200 pads or interconnections on a single chip, and we had over 750 pad chips in production. And so over the years C-4 has proved to be an extremely good technology. We did a lot of experiments on it, because as we kept extending the I/O count we always wondered if the technology would reach its limits. However, we were able to achieve a 99.95% manufacturing yield on joining the chip almost independent of the number of balls on it, at least up to 800 pads (where I retired). We’re not talking about 99.95% yield by the pad; we talking about it for the chip, which is a really great manufacturing process. Those were a few of the interesting things.
Ferrite memory cores, semiconductor memory
Another thing that was interesting - when I joined IBM my first job was testing ferrite memory cores. Up until 1968 -69, computer memory was dependent on ferrite core technology. When I started in 1957, the ferrite core was 80 thousandths in diameter with a 50 thousandths hole in it, and we put three wires through that. You would have an X wire, Y wire, and a sensing wire Z, so you had to get three wires through every core. When I was re-assigned to passive components in about 1959, we were working on a 0.010 inch diameter core with a 0.006 inch hole, still trying to put three wires through it, and it was getting very tricky. By 1964-65, core memories were being made with 4000 bits - a lot of memory in those days.
It was. I remember learning FORTRAN on a 1620, which was still core memory.
Yes. And along about 1965, we made our first semiconductor memory, and it was a 16-bit chip. That went into production, but only in a small, small way, for a single system, just to get our feet wet making it. And our first real main memory came out with a System 370. I was the packaging manager for that program. We put four 32-bit chips on a single half-inch substrate, and then we stacked substrates so that we could get eight chips in a half inch area. That was if you had all good chips. In those days you didn’t always have all good chips, but with eight chips to play with, you could get at least four equivalent good chips. So we played a lot of games with partially good sectors on the chips, as we called them, to get the yield up. Yields in those days on those kinds of chips were not great. We were processing 2¼-inch wafers and later went to 3-inch, but when we first started making those if you got one or more good chips per wafer, you were over the moon. So that was really pushing the state of the art. And of course the semiconductor main memory has gone from there. Now those were bipolar memory chips.
MOS memory, managing the thin film substrate program
We switched to MOS memory along about 1970-71, with a big step up to 128 bits per chip. So the MOS memory chip had more capacity and it also had lower power consumption and was easier to cool. Meanwhile, we kept expanding the hybrid capability, cranking things down tighter and tighter in tolerance, going to higher densities of wiring. We pushed the state of the art on screening technology very hard. We found that 0.004 inch lines and spaces were about the best you could do with screening technology. There was just not enough metal in the inks (electrode pastes) to do any better than that. And, I believe, that seems to still be pretty true today. Along about 1971 we started moving into thin film substrates and in 1974 I transferred up to Endicott, NY as manager of a thin film substrate program, which we called metallized ceramic. We used a copper thin film—thin meaning about 80,000 Å of copper, about 3/10ths of a mil, so it’s not terribly thin. We evaporated or sputtered the copper and formed the patterns with photolithography. This allowed us to make two mil lines and spaces rather readily.
Meanwhile, the industry in the late sixties also moved into integrated circuits, and you could buy five to ten circuit chips in plastic packages. I think our procurement people spent a lot of time with Texas Instruments on the molding compounds, because pure molding compounds in those days were nonexistent. So it was a really special job to get a molding compound that would be pure enough for the integrated circuit. I remember that they ended up with the Novalac epoxies as being the prime material that went into very large production. Epoxies, in those days, were relatively “dirty”, meaning they contained free chlorine and things that, in the presence of moisture, would give you a number of reliability issues.
Thus the importance of all the chemical engineers in the packaging industry.
Yes. Epoxies were much less expensive than hermetic sealed packages, so a lot of work went into them. But the volumes of epoxy used were not always very large. Take my metallized ceramic program (since it is one I’m familiar with), we used a back seal on a 1-inch substrate. We were building about 50 million of these a year. Well, every year I had to go out and beg 3M to do another small pilot run for my back seal material, as it was one they formulated especially for us. I had to buy a year’s supply at a time; they just did one run a year. They made a few hundred gallons, and I stored them in a warehouse until I used them up or the shelf life ran out, and then I’d go back and beg them for more. Basically a few hundred gallons is just a pilot line run, and it didn’t pay them to put a lot of effort into these things, because you just don’t use enough of it. So we to struggle in the early days, in the late sixties and early seventies, to get really clean molding compounds from anybody. There just wasn’t enough volume business to attract a lot of development work. Of course it became very important to the industry, so over time it became a good business.
CPMT Society involvement
I’d like to ask about your involvement with the IEEE CPMT Society and how you feel the CPMT or its antecedent societies assisted the industry, and how it worked with the industry and grew with the industry.
I got involved in CPMT kind of casually in the mid-seventies by giving a paper at ECTC and then liking the conference. Some friends of mine, people I knew at IBM, were active in the conference—Diana Bendz, who is still at IBM; John Powers, who after he left IBM was head of IEEE for a while and is now a consultant. I enjoyed coming to the conference. It was a good conference to be at, and there were a lot of things in which I had an interest. Along about 1979, I believe John Powers was president of CPMT at the time, and he was pretty unhappy with what was happening in the packaging side, so he begged me to take over TC-6, which was the packaging technical committee. He felt that there was a lot there that could be done. I called up the former chair and asked him for a list of people that had been working in the packaging. He said basically there wasn’t anybody; you can’t do packaging in CPMT, there is no interest, and “that’s why I quit.” So I started with no committee, no nothing, and we kind of went from there. George Harmon over at NIST (National Institute of Standards and Technology) was a big help. He and I sat down and did some brainstorming about how to get this going, and the first thing we tried to do was to put together a couple of sessions for ECTC on packaging. We did that in 1980. And then George suggested starting our own VLSI packaging workshop, and so we did that, and I chaired that with George’s help, and that was a big success. I don’t remember whether that first one was in ‘80 or ‘81, but we had a good turnout, a lot of interest, and suddenly the packaging interest started taking off. So then in 1981 we went out and collected so many papers for ECTC that I came in to the program chair and said, “I’d like to do a minimum of three sessions”. Well, the other committee chairs didn’t like that at all. That was hogging the conference, and that was no good. But the program chair said, “Wonderful idea.” And so we had three good sessions on packaging, and it just kept building. From ‘82 to ‘84 I was on assignment in England for IBM, but I did come back to ECTC each year and I kept finding papers, and started soliciting them in Europe. When I came back in ‘84, I was asked to become assistant program chair and start through the chairs for the conference. When I was a program chair, I set a new record for getting abstracts for the conference. That was funny, because there was something like 166 papers. We only used 96 papers in those days. And now we’ve expanded with the times to attract more papers. After I left, C. P. Wong became program chair and he beat my paper record handily, and the conference has just kept growing ever since. But the conference had to change, because companies started saying, “You can’t go to a conference unless you’re giving a paper.” And so instead of limiting the number of papers, we had to try to find a way to get more authors. The solution was to expand the number of sessions and run more sessions concurrently. . We’ve also made provisions for poster papers, and then started providing short courses. Both of these things have worked very, very well. It has been a big draw. The success of the short courses is due to our vice president of education. He did a very simple thing. He said, “Every year I’m going to take the two bottom courses and throw them out, and bring in two new ones.” And so by providing that turnover, the program keeps right up with the times, and that’s worked marvelously well.
CPMT leadership roles
But that’s how I started into CPMT. Then of course one thing leads to another, and the next thing I know I’m running for the board of governors, and I got elected to the board. Then they asked me to be administrative vice president in ‘88. I was administrative vice president for two years. There is quite a bit of turnover. Normally in those days the technical vice president would run for president. It didn’t work that way, and in ‘89 I was asked to run for president for the ‘90-‘91 term, and so I became president of the CPMT. The CPMT never lets anybody off the hook. After you’re president, you become chairman of the finance committee, and two years later you become chairman of the strategic planning committee. And so as past presidents, they keep things moving along. CPMT per se has one of the best organizations of any society in TAB [Technical Activities Board (IEEE)]. In about the mid-eighties, the board of governors commissioned John Powers (as a past president), to put together a strategic plan for CPMT. He did a really great job of that, and I still have a copy of that somewhere. He laid out some of the challenges, and he laid out an organizational structure that established technical and administrative vice presidents, and a committee structure that would report to them. He also provided for a vice president for publications, and put a framework in place that could be extending that structure. He appointed all the committees and who they would report to and how we would manage. We followed that organizational structure pretty faithfully, and we have quite an excellent organization at the committee level, and then assignments for each of the committee chairs, and vice presidents who oversee particular sections. We added a vice president for education, and we added a vice president for conferences. And we did a couple of unusual things. The reason I say this, is when I was president we started the TAB Review Committee. The TAB vice president at the time later became the president of IEEE, Troy Nagel. I think Troy was trying to shake up the old establishment to some extent, and of course with 36 new society presidents, we were all behind him and felt the new blood and shaking the tree was a good idea. He started some things like society review committees. You know, some of the societies had never had a review. The bylaws say we should, and so we started a committee under Bob Begun, who was president of the Power Society at the time, I believe. I started looking into it, because I was trying to find ways to improve CPMT and thought maybe I would get some ideas from what the other societies were doing. Bob had kind of drafted out a little plan, and I marked it all up and gave it back to him with a lot of comments. He liked that so much that I was appointed to the committee. So over the next four or five years I sat through 10 or 12 society reviews, and I found that none of the other societies have as good a management structure as we do (My opinion, of course). CPMT has a high percentage of practicing engineers as members and tends to be very practical. Troy at the time thought that was great. He was encouraging us to have more practical management and papers in the societies as opposed to the theoretical academia approach. Today it’s kind of swung around. Our president today tells me he is getting pressure to put more academics on the board, and we’re all saying, “Don’t do it.” This tends to be a fairly practical society.
Anyway, as president of CPMT I was surprised at what I found. I found that we were one of the best managed societies in TAB, and that we had a lot of good things going for us. The other thing I found was that most of the societies have a great deal of frustration with their technical committees, claiming they didn’t do anything and they weren’t pulling their weight, and they couldn’t get out of them what they expected. I’m sitting there saying, “Gee, our technical committees are great.” What is the key difference? Well, the one key difference was that whenever we run or a workshop or a conference, the chairman of the technical committee in that field had to endorse it and support it, and it becomes his or her conference. So the technical committee has the responsibility for a workshop or a conference, and that means that they’ve got something concrete that they have to do, that they have to work on. And of course we also pull them together every once in a while to talk about strategies and where things are going and so on. But that little trick of giving them responsibility for conferences got them involved with something meaningful to do that was visibly measured. And if they didn’t do that, we knew to replace them with somebody who would. But no other society that I am aware of does that. And that was the main difference that I could see in what we do that the other societies don’t.
One of the things that annoyed me while I was president was our awards plan. We had one official award and we had one unofficial award, which was manufacturing technology. We had been trying to do something with manufacturing technology. The awards chairman said, “I’ve got too much to do anyway, and I’ve had this for too long, so how about you take it over?” So I did. I took over chairman of the awards committee, and I pulled together a group and we sat down and brainstormed and came up with several new awards that we took through TAB and got approved. And basically we went from just a contributions award to an outstanding sustained technical contribution award, an additional one; a manufacturing technology award; and a young engineer award. So we expanded our formal awards from one to four, and as the Transactions expanded, we expanded the Transactions “best paper” awards as well. I think there is still more that could be done, but at least we got that far and got some new things on the table. The thing that we don’t do yet with our awards is publicize them adequately, nor choose highly visible recipients. We are good engineers but very poor when it comes to public relations and politics.
Which is a good thing.
Which is the good thing. The bad thing is that we get these awards out and we get little or no publicity or recognition for having them. I was after the awards board to not only look at the guys who are coming in to be nominated for the awards, but to keep watching for some high profile people that might be nominated for the same award that will make a splash about CPMT when the award is granted. Not that I want to give them to someone undeserving, but if I’ve got two people of comparable worth and one is high profile, I’d rather get the publicity for the society. We still don’t do that. We still have a struggle trying to get the awards nominations and pick the right people for it. You’ve got to do a little bit every time. Every year we do a little bit better and we continually improve, that’s what’s important.
CPMT chapter, Binghamton
But then I retired from IBM in ‘92. Although I was doing some consulting work, I didn’t have enough to do, so I formed a CPMT chapter in Binghamton. With the help of Don Seraphim, an ex-IBM Fellow retiree, we decided we ought to run a workshop. So I went to the board and said, “We have a new chapter and the VLSI packaging workshop is in Japan this year, but we badly need an area array packaging workshop. How about if we in the Binghamton chapter run it in the fall with a 50/50 financial advantage with the society? What we want is for CPMT to loan us $5,000 and we’ll run the conference and pay you back and we’ll split whatever surplus there is.” And I told them, “Here’s my budget,” and I showed the 15% surplus, which is our rule of thumb on our conferences, and how we would achieve that. The board of governors went along with it and gave me the advance money, and we started putting together this conference. We realized we didn’t have a lot of time, because normally you like to have a year to plan these conferences and we had about six months or less. So we put a little committee together and went out and started soliciting papers around the main theme, and we put together a whole conference based on just this one theme. We got a lot of up-to-date and good, practical papers. Now this is Binghamton, New York and, although the hotel was only going to cost $58 per night, we’re upstate, and not that easy for people to get to. We put together the whole conference, and based it on getting 75 people to attend. We thought that was pretty realistic, for attracting people to come to Binghamton. In planning the budget, we used normal expenses. But in Binghamton, we found the hotel meals rather than being $40 a dinner was closer to $20 and so on, and we put together a little show for one of the nights. We decided to do a kind of a get together with the group on the first night of the conference to break the ice and kind of get everybody working together and talking. So we found a mansion not too far from the hotel in Binghamton that we could rent for the evening and we could have a catered dinner, and arranged that they would take us on tours of the mansion—the Phelps Mansion, which was built about 1890, with beautiful woodworking and wainscoting. Then we worked out the fee for the conference, and that fee came out to about $250 a head. That was all we needed to charge. In 1994 that was still pretty cheap! I mean, people were paying $400 for comparable conference arrangements. The numbers worked out, so that’s what we did. And then we had a number of big companies in Binghamton, including IBM, Universal Instruments, Lockheed-Martin, and we had people that were working for those companies that were part of the IEEE chapter. You know, companies have a big thing about sending people to conferences, but this was right here in town, so I said, “How about this offer? If you want to take at least ten seats, I’ll give you ten 1-day passes at a cut rate and you can have your people come in for the day or for a session, and it doesn’t cost you much. They can just drive over to the hotel, listen and go back. If you guys would like to do that, we’ll arrange it.” I ended up selling about 70 of those. Then the conference people started calling in and registering. Three weeks before the conference we had to stop taking registrations, because the room would only hold 150 people and we already had 170 registered. They couldn’t give us the other half of the ballroom at the hotel, because there was a wedding in there. So the nub of all that, when we were all said and done, was that our surplus instead of being a couple thousand dollars ended up being closer to $18,000. So we paid the society back and then gave them another check for $9,000 on top of that.
They must have been very pleased.
They were very pleased. We ran it again in ‘96, and we tried not to make so much money and we didn’t charge that much, but we still cleared $15,000. When I left Binghamton the end of ‘96 to go down to Ashburn, Virginia and work for NEMI, I left the richest chapter in the Section. The Section just couldn’t believe that we could come in and run a conference and make that much money. Their budget was between $3,000 & $4,000 per year, and most of that came from RAB. They had really no fund raising of their own.
No corporate backers?
No corporate backers. All IEEE members are individuals, not corporations.
<flashmp3>277 - gedney - clip 3.mp3</flashmp3>
If you tap the corporations for backing, then you’re got to listen to some of the things that they want you to do, and as a professional society you may find yourself kind of boxed in. If you want to promote portable pensions and your big company members don’t want to, how do you handle that? So I think we’re probably better off not to have the corporate backing. But there are times when the money would be nice.
CPMT secretary role, budget
The current CPMT president asked me if I would become secretary to the CPMT so that I would have an excuse to keep attending the board meetings. I’ve done that for the three or four years, but really for me there is no work. Marsha Tickman does all the work. This is fortunate for me, because with my hearing problem, I can’t always pick up everything that goes on. She records it and then goes through and pretty well writes the minutes, and she is very, very efficient. Being secretary means I show up at the board meetings, plus what Marsha does. Another interesting thing is that two years before I became president of CPMT, Leo Feinstein was president, and the board charged him with taking on some rather significant new endeavors, which was to support Circuits and Devices Magazine, support a new manufacturing technology conference, and to increase Transactions pages. The net result of those things was that our reserves started going down hill extremely rapidly. Back in 1989, our treasurer quit because somebody criticized her, and whenever you criticize a volunteer they go, “Goodbye.” Well, she quit in the middle of the budget cycle, and we did not increase our Transactions subscription charges, in the budget, overestimated income and underestimated costs. As a result we were sliding downhill at the rate of $60,000-$80,000 a year, and our reserves in ‘88 were only a quarter of a million. So in 1991, my second year as president of CPMT, our reserve went as low as $60,000. In 1990 I could see this was happening, so I started working on turning it around. It took me almost a year to convince the board that we had to do some drastic things. Boards do not swing around easily. The other thing we had agreed to was to take on a scholarship in connector technology with the understanding that the Holmes conference would generate enough surplus to pay for it. But then they had a couple of bad years and ended up with a deficit. The manufacturing technology conference cost us $20,000 a year three years in a row, and Circuits and Devices Magazine was running between $20,000-$30,000 a year, which was money out of our pocket we couldn’t recover. So, first we dropped Circuits and Devices Magazine. We had two past presidents and a vice president as a steering committee for that the Manufacturing Technology conference and they were big money losers. Leo Feinstein as president made a number of suggestions, and then I made a number of suggestions that first year of how that conference could be gotten back into the black. Basically what it was is they were getting 150 people to a conference where they were planning for 250, so that their expenditures were far outstripping their revenue. My solution was to either get together with another conference and share the venue and expenses, or start planning for an attendance of 150 at conferences. The committee basically ignored Leo and ignored me. So I sent them a letter, and they all resigned. But they came to meet with me. They said, “Obviously we’re not doing what you want,” and I said, “That’s right.” And they said, “We don’t want the conference to go down the tubes,” and I said, “Okay.” And they said, “So we’re going to tell you everything we know about it and help you make the transition,” which was really very good of them. Basically we pulled a new committee together, and I told them that they had one year to either make it pay or we were going to cut it out. So they put together a budget, and that year they broke even, and the next year they started making money. It was manageable to do. So that’s how we got out of that one. Though we dropped Circuits and Devices Magazine, we had to pay for another year, because we had to give them a year’s notice. We also dropped the connector scholarship, and generally tightened the reigns on spending. We increased our subscriber prices on all our Transactions, and went back to TAB with a special presentation to show them what had happened and why we were losing so much money, and asked for a special dispensation to increase prices by 40%, and they let us do that. And then I started holding the vice president of publications responsible for meeting his page budget, because he had overrun his page budget two or three years in a row. Between those actions, in 1991 we showed a slight surplus. Not much, but we got a reserve up above $80,000. Then fortunately the president after me, C.P. Wong, stuck right to it, and then after him Dennis Olsen did as well. We pulled out, and we’re in good shape today. We’ve got a surplus today about $1.5 million, over 100% of our budget. We’ve done very well.
Entry into packaging and manufacturing field
It was very nice talking to you, and I’m very glad to get this history. I do have one question. Back in the beginning, what led you into the field of packaging and manufacturing?
When I joined IBM, fresh out of college, my assignment was working with vendor componentry, qualifying it for IBM systems. I worked in this area for about 8 years, before moving into the development laboratory where I started as an assistant to a senior manager, and was exposed to a number of new technologies. Silicon designers were very scarce at the time, so several of the packaging managers with device background were strong-armed to take jobs in silicon design. I was picked to backfill and given a group charged with developing ceramic packages for IBM memory chips. I found that the process development side was extremely interesting and challenging, and enjoyed the work.
Did this go back to things you had studied in school? What sort of preparation had you received?
Well, if you have get an engineering degree you tend to study a little bit of everything. I just found it very challenging, all the materials problems and physics problems along with some of the electrical problems. My job primarily over the years was more chemistry and physics than it was electronic. But having an understanding and appreciation for the electronics side was very valuable in working with the system designers and helping me to be able to judge if what we were doing was right. I am probably a little unique in that not too many electrical engineers manage those types of areas. After I got out of college and went to IBM, I started working on a masters degree in physics. I found that I was more interested in the physics than I was in digital circuit theory. That probably helped, though I never completed the masters degree. I was two courses short when I was put into management. The next semester I found that after a 10-hour day at work I was falling asleep in class. Management just carried too much responsibility with it. But basically, I got interested, and I just found process development as a whole to be very challenging and interesting, with enough of an edge that I could see what was going on and understand it all. Although I did not have the technical depth that I could adjustment the chemistries or that sort of thing by myself, but I could judge what other people were doing in terms of the management side. It was quite a challenge, and very enjoyable. I also enjoyed releasing products to manufacturing. There is great satisfaction in seeing something you have designed go into volume production. I was responsible for the release of literally hundreds of part numbers, and there are some things you have to learn to be able to do that. First knack is knowing when to stop development, freeze your design and release it. Second is to be able to prove that it will go into production and work – do enough work and gather enough data to prove it out. Third is in support manufacturing. You don’t have to exactly hold their hand, but you can’t stand back and watch them fail with your product. You have to make sure it succeeds. If the manufacturing team gets into trouble and you’re not there to help him, he resents that, and he’s not going to take the next release from you. But if he’s got a problem and you go in and help him fix it and give him the credit, he is going to be willing to take the next release from you. I always tried to maintain a good relationship with my manufacturing counterpart, and I always tried to anticipate what problems they were going to have while I was doing the development work, so that it would go in smoothly.
Well, that’s been very, very helpful to me, and it is really excellent background. Thank you very much for all your help.
- 1 About Ron Gedney
- 2 About the Interview
- 3 Copyright Statement
- 4 Interview
- 4.1 IBM, IEEE, and the computer industry, 1950s-1960s
- 4.2 IBM Components Division
- 4.3 Underestimation of computer market
- 4.4 Component suppliers
- 4.5 Reliability, testing, and standards
- 4.6 Screened paste resistor technology, substrate technology
- 4.7 Controlled Collapse Chip Connection (C-4) technology
- 4.8 Ferrite memory cores, semiconductor memory
- 4.9 MOS memory, managing the thin film substrate program
- 4.10 Packaging industry
- 4.11 IEEE activities
- 4.12 Entry into packaging and manufacturing field