From GHN(Difference between revisions)
|Line 581:||Line 581:|
Going way back to an even earlier inspiration was a man named Harald Friis. He was a radio pioneer, a great engineer and a great human being. He had a tremendous influence and inspired a whole generation of radio engineers and radio designers.
Going way back to an even earlier inspiration was a man named Harald Friis. He was a radio pioneer, a great engineer and a great human being. He had a tremendous influence and inspired a whole generation of radio engineers and radio designers.
Revision as of 16:06, 26 January 2009
About Eugene O'Neill
O’Neill received his bachelors in engineering in 1940 and his masters of electrical engineering in 1941, both from Columbia University. He then worked at Bell Labs for the rest of his career. During World War II he worked on airborne radar systems. After the war he worked on the L-3 coaxial system until the early 1950s. He then worked on microwave radio systems; transoceanic submarine cables in the mid 1950s the time-assignment speech interpolation terminal system (TASI) in the late 1950s, which increased capacity by using high-speed switching to use a channel only when a person is talking; and the Telstar project, an early communications satellite, in the early 1960s. Then as executive director in charge of all long-distance transmission systems, he developed the production version of short-haul digital transmission in metropolitan areas. He also developed the L-4 and L-5 coaxial systems, and the single sideband system. He is most satisfied with his work on TASI, followed by his work on metropolitan digital terminals. Telstar was nice, but he was too busy administering, too far from the actual research, to feel as much satisfaction about that.
About the Interview
EUGENE O’NEILL: An Interview Conducted by David Hochfelder, IEEE History Center, 9 July 2001
Interview # 415 for the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc., and Rutgers, The State University of New Jersey
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, Rutgers - the State University, 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:
Eugene O’Neill, an oral history conducted in 2001 by David Hochfelder, IEEE History Center, Rutgers University, New Brunswick, NJ, USA.
Interview: Eugene O’Neill
Interviewer: David Hochfelder
Date: 9 July 2001
Place: Middletown, New Jersey
Let’s start by talking about your college education.
I went immediately into Columbia University after graduating from New York City high schools with the intention of following a course in engineering. Columbia had no freshman engineering classes. I took a pre-engineering course in the Liberal Arts College. While I think that slighted the engineering side of things a bit, I have always been grateful to have gotten a better background in history and English than I would otherwise have gotten. In the third year students pursuing engineering were transferred from the Liberal Arts College into the Engineering School. I did well there and got my bachelor’s degree in 1940. Following that, I went to work for Bell Laboratories for one summer. I became quite taken with the place as a prospective employer, not least because I was a New Yorker and the idea of working in New York appealed to me.
I returned to Columbia for a fifth year, at the end of which I got a master’s degree in Electrical Engineering with specialization in Communications. After graduating from Columbia University in June, 1941, I immediately went to work as a new employee at Bell Laboratories.
Bell Labs employment
World War II, airborne radars
What was your initial work assignment at Bell Labs?
It wasn’t what I had wanted. I was full of the lure of electronics, vacuum tubes, radio and those sorts of things. However the only job I was offered at Bell Laboratories was in switching. In those days switching meant clinking and clanking relays. There were no vacuum tubes in sight and it was pretty low frequency. I remember my reasoning very well, figuring that the thing to do was to get a job at Bell Laboratories and then in the future there would probably be maneuvering room to get into something there that was closer to my heart. In those days they had an apprenticeship program. One started out by adjusting relays and wiring frames and so on.
Just a matter of months later, on December 7th, 1941 the whole complexion of things changed with the attack on Pearl Harbor. I was immediately sought by several different parts of Bell Laboratories and I elected to work on a 3-centimeter airborne radar development. Within a few weeks after December the 7th I was working on airborne radars there.
Did you work on airborne radar throughout the war?
Except for minor deviations, yes. We worked on some specialized radars for mortar battery locations that didn’t amount to much. Essentially the entire war was spent working on airborne radars.
Post-World War II projects, microwave radio relay telephony
What sort of work did you do after the war?
At the end of the war Bell Labs resumed commercial non-military development.
To backtrack a little bit, the first coaxial cable system with a bandwidth of about 1 megahertz was completed just before the war. Some transmissions had been done between New York and Philadelphia. Also, by the end of the war television was looming. There was a tremendous pent up need for high capacity long distance communications. Bell Labs immediately started work on microwave radio relay telephony. That was stimulated by the radar development during the war. They also started work on the next generation of coaxial systems. My group landed on the so-called L-3 coaxial system, which was a 3-megahertz system. Our objective was to get a bandwidth broad enough to transmit 600 channels of telephony. I worked on that from late 1945 or early 1946 until its initial installation in the early 1950s.
Was this system also used for television transmission?
Yes, it was. In fact, that was one of the prime motivations for it. It was not just higher telephone capacity, though that was very important to us. We also wanted to transmit the television signals of the day. We tried very hard. We pretty much lost out in that endeavor. I do not think they transmitted much television on a commercial basis, though we were credited with that. When there were problems in television transmission in the early days it was often blamed on coaxial cable, but if truth be told, far more television was transmitted over the microwave radio relay systems. There was some use of coaxial cable for television, but not a great deal.
Transatlantic submarine cables, long distance microwave radio transmission
You had some involvement with the submarine telephony cables too. Is that right?
Yes. The development of the very first transatlantic submarine cables was in the laboratory immediately adjacent – above or below – to the laboratory in which I worked on coaxial cable. They did some testing where trial cable systems came down through holes in the ceiling. In the laboratory where I was working these submarine cables would occasionally snake down through our laboratory and disappear through the floor into the laboratory below. People working on the submarine cable were good friends of mine, but I was not personally involved in the first transatlantic submarine cable.
After the development of the 3-megahertz L-3 coaxial system there was a brief interlude where I switched to microwave radio. I worked on a project called the TJ microwave radio system. That was an 11-gigahertz system. The objective was for a low-cost skinny route system and not a main long distance system. The concept was that it would be used for spurs off the main microwave or coaxial routes to connect outlying points that were not directly on the main long distance routes. It went into manufacture and we used small amounts of it in New Jersey and elsewhere, but it never really became a significant factor in microwave radio. Short haul microwave radio got a lot of attention in Bell Laboratories, but we never had any great success with the short haul versions. Apparently the costs were too high.
Was there more success with the long haul versions?
Microwave radio exploded in the long distance business. The first experimental microwave route– in which I was not involved – went from New York to Boston and was quite successful. That must have been about 1947. By 1951 the first transcontinental microwave radio system was completed. Following that, microwave radio transmission expanded rapidly. There were a lot of coaxial installations too, but microwave radio amounted to about two-thirds of all the route and channel miles that were installed. That continued throughout the 1950s.
The first telephone transoceanic submarine cables were put in service in 1956 from the United States to England. These were really the first high-quality long distance transoceanic telephone circuits. Before that, starting back in the late 1920s, we had voice circuits via high frequency radio, but it was full of fades and static crashes. It was very difficult to understand speech through that medium. However it grew to a pretty considerable worldwide network since it was the only means available. The so-called TAT-1, 1956 transatlantic submarine cable, provided thirty-six voice channels multiplexed on the cable and was high quality. The quality was entirely up to the standards of domestic long distance transmission. Traffic took off with a great rush. That was in 1956.
Time-assignment speech interpolation terminal system (TASI) project
In 1957 I got on what I still consider one of the best projects with which I have ever been associated: the time-assignment speech interpolation terminal system (TASI).
When I was assigned to the TASI project, I had already been promoted to department head and had an excellent group of about twenty-five people. The project was fascinating in every respect. First of all, it was all solid state. Transistors had been invented in the very late 1940s. They began to be commercially applicable in the late 1950s.
Here is an anecdote. No one disputes that the transistor was invented at Bell Laboratories, but the statement is often made that Bell Labs didn’t know what to do with the transistor until the Japanese picked it up and made little cigarette case sized radios out of it. That is a wrong depiction of what actually happened. We knew the potential and we were very eager to use solid-state devices and transistors, however it was not until the late 1950s that devices of the reliability absolutely necessary for applications in the telephone plant became feasible.
This time-assignment speech interpolation was one of the first applications. It was not only all solid state – the first solid-state devices that I had worked on – but it was almost completely digital. It had stores, circulating memories and speech processing of various kinds. It was an absolutely beautiful project with tremendous fellows with whom to work. We also had the glamour of having terminals in New York, Paris, London and Hawaii. That certainly did not lessen its appeal. I commuted. I was in charge of the project, but had associates who had the assignment of spending six or seven months in Paris while the terminals were installed there.
Would you explain in technical terms how the TASI system worked?
To this day, long distance circuits have one path for transmission one direction and a completely separate path for transmission in the opposite direction. Transmission does not go over a single pair of copper wires or a single optical fiber. Transmission from east to west is over one wire, coaxial, radio channel or fiber and transmission from west to east is on a different wire, coaxial, radio channel or fiber. That being the case, in the course of normal conversation one person’s half of this two-way split circuit is busy less than half the time. In most normal conversations one person talks while the other person listens. While one is listening, that person’s side of the circuit is idle. The idea of the time-assignment speech interpolation was to connect someone to a channel only when he was actively speaking.
On the first submarine cable we had thirty-six full-time working channels in both directions. As an example of the two-way concept, there were two separate submarine cables in the first transatlantic system – an eastbound cable and a westbound cable. We built terminals equipped with speech detectors that would recognize which of the thirty-six channels was active and someone was talking on it. This permitted about seventy-two people to talk on the thirty-six channels. When a person stopped talking we disconnected him momentarily and assigned his channel to a speaker who became active at that time.
To enlarge on this a bit further, speech activity is considerably less than 50 percent. This is because one not only pauses to listen to the other person, but also because in the normal course of conversation even when speaking there are gaps, pauses and spaces. When people talk on a telephone line, they are actually sending out electrical signals perhaps about 35 percent of the time. Statistically, if a channel is occupied only 35 percent of the time and there are thirty-six channels, then we estimated that twice that many speakers could be accommodated by assigning them to those idle intervals. The switches must be very fast, with very quick detection of speech activity.
The far end was signaled to let it know which speaker was assigned to which channel. We had to send signaling signals as well as speech signals in order to convey the switching intelligence. In effect, the TASI terminals were high-speed switches to connect a larger number of speakers – say seventy-two-to thirty-six channels when any thirty-six or fewer of them were speaking simultaneously. That is what TASI did, and it was a great success. It was a lovely project. Traffic on the cables had grown so rapidly that it was something of a crash project. This interpolation scheme did double the capacity of the cable. I remember it fondly as a very great period.
What were some of the technical challenges involved in the TASI system?
It was all new ground for us. We were not used to digital technology and had to learn a great deal about that. Nor were we used to solid-state implementation. The terminal to interpolate seventy-two speakers onto thirty-six channels occupied ten or twelve large bays of equipment. We had no integrated circuits. There were only discrete components with transistors and diodes by the thousands.
One of the great challenges – in addition to the basic technical problem of developing sensitive speech detectors, very high-speed signaling, and very high-speed switching equipment and things of this sort – was reliability. Solid-state electronics was the godsend. We could never have been built this around perishable vacuum tubes. With the reliability that the solid-state components brought, we were able to carry it off. Even then we had some tough moments. Nowadays solid state is almost a synonym for high reliability. We had thousands of components and it didn’t take much of a weakness in one of them to cripple things. We had a few of those kinds of problems to overcome, but we did.
Was this Bell Labs’ first foray into digital equipment?
We like to think so. There are people who would disagree. It certainly was one of the very first. There was an all-digital private branch exchange that was solid state. I would not claim – because I don’t know – that TASI was the first digital solid-state development at Bell Labs. Perhaps it was not. It was certainly among the first very few and pioneering in that regard.
There were predecessors to TASI. There were earlier ventures and quite a lot of research work on speech interpolation, but they were experimental setups in laboratories. Looking back on it, none of them really embodied all the features necessary to make it work. There was at least one – or maybe two or three – key inventions in connection with the TASI development that made the thing practical. The earlier efforts were not practical and one could never have put them into production. In the late fifties the United States was trying – and as often as not failing – to put satellites into orbit. Everyone assumed that those difficulties would be overcome and satellites would be put into orbit.
ECHO experiment with NASA
It was in this general atmosphere that the laboratories proposed, and NASA agreed, to cooperate in the so-called ECHO experiment. As I understand it – and I am not expert in this part of the history – the concept of putting a large inflated sphere in orbit was originally NASA’s. I understood that it was to be an experiment in the physics of the very thin upper atmosphere, essentially outer space, and this sphere would be sensitive to the presence of and react to air molecules.
Two very bright and inventive fellows at Bell Laboratories, John Pierce and Rudi Kompfner, got wind of this attempt by NASA. They went to NASA and said, “If you are going to do this, this has some very interesting possibilities as a communications satellite. The sphere needs to be aluminized so that it will reflect radio waves.” NASA was receptive to that. Working with Goldstone, the government space and antenna installation in California, they built a special antenna on Crawford Hill. That’s just down the road from where we are sitting now. The satellite was a 100-foot diameter aluminized balloon. There may have been one attempt that failed, but they finally got the shiny mirror surface sphere in orbit in 1960.
The attempt to transmit signals to and from Goldstone, California to Crawford Hill in Holmdel, New Jersey was successful. They were able to transmit voice and still pictures. They were not able to transmit television, i.e., full motion video, but it showed that it could be done. One could transmit up to a satellite and then by re-radiation from the satellite one could receive signals at an acceptable level. There were many of developments that contributed to this. This was right on the heels of the development of masers – microwave amplification by stimulated emission of radiation. These were ultra-sensitive, very low noise radio receivers. By virtue of that and some very low noise antennas they were able to transmit via the ECHO satellite.
As soon as ECHO was demonstrated, people realized the great potential and made many proposals for satellites. Even before ECHO John Pierce made a proposal as early as 1955 and Arthur Clark made an even earlier proposal in the 1940s that, if someone could make an active satellite, one which would receive a signal, amplify it and retransmit it – that this would be an enormously useful thing. As soon as the passive ECHO satellite proved to be successful, there was tremendous excitement in anticipation of building such an active satellite. That was the birth of the Telstar project.
At the time of the completion and success of the TASI terminals I was working for A. C. Dickieson. He died just recently, by the way. He was my boss when I was working on TASI terminals. He asked me to head up the development of the Bell Labs version of the active satellite, the Telstar project. I got onto that in 1960.
What were some of the technical challenges involved in Telstar?
How could I list them? We take satellites and communication to and from satellites so much for granted now that it is easy to forget some of the uncertainties that were in people’s minds at that time. I have already mentioned the ultra-low noise maser receivers. They operated with noise temperatures and sensitivities way beyond anything that was previously possible. That was a very big step. Without that, satellite communications would not have been practical at that time. Nowadays people have 18-inch satellite dishes in their backyards. Now we have much larger and much more powerful transmitters in the satellites than we had any prospect of having at that time. In addition no one was absolutely sure that transmission up to a satellite and back down would be essentially free space transmission.
ECHO did a lot to reassure us on that score, but there were phenomena like absorption by water vapor in the atmosphere, so-called Faraday rotation of polarization and other uncertainties about the real transmission environment. We worried a great deal about those things.
There was no space shuttle. Once this thing was out in space there was no repairing it. That was another concern. Integrated circuits were in their infancy and barely existed in 1960, and there were essentially none with the qualities we needed available. Nor did we see any prospect of having them in the very near future. Reliability was a tremendous worry. Submarine cables are installed in a rugged environment, and laying a submarine cable and active amplifiers from a cable ship is a pretty grueling physical test of the equipment. However, launching a satellite by a rocket from Cape Canaveral takes the test to another level. The physical requirements for ruggedness and vibration resistance and things of that sort were something that we had never coped with before. That was a major concern and we went to great lengths to try to address those problems. After we had assembled the Telstar satellite and satisfied ourselves it was working, we filled the container holding all the electronics with foam plastic. That was to keep the pieces from being shaken to destruction during rocket launch.
There was a big disagreement regarding the frequencies. The ground receivers were ultra sensitive and we were not quite sure where we could locate ground receiving stations. We were unsure what frequencies would be allocated and that that we would not run into interference from the widespread land-based microwave systems which already existed by 1960. It was not then, and has never since been, any picnic to get radio frequencies allocated. I don’t have to tell any radio engineer that. Our conclusion was that the only frequencies that made any sense were the existing common carrier frequencies in the 4- and 6-gigahertz bands. There is allocated about 10 percent bandwidth in each of those two bands. We decided early on that those were the appropriate frequencies to use, but these were the same frequencies that were already being widely used in terrestrial microwave systems. The potential problem of interference into ground station receivers was very worrisome.
There is a footnote to that. Our early concept was not the familiar geostationary satellites which became the standard in communications until fairly recently. We pictured low-altitude orbit satellites in fairly large numbers that had to be tracked from horizon to horizon. When the one in use set transmission had to be handed over to another still in view. The antennas were not fixed in orientation. They were tracking the satellites. That made them especially vulnerable to ground-based microwave radio system interference. That was a big challenge.
In retrospect, I think maybe our concerns were overstated. Radio interference turned out not to be as difficult a problem as we thought. However it did lead us to locate the ground station in a remote corner of northern Maine.
We were thinking in terms of low-altitude satellites. Therefore the period of mutual visibility – when the satellite could be seen from both Europe and the United States – was limited. As a result, we thought our station should be as close to the east coast and as far north as possible. The British put their antenna in Cornwall, as far west as they could get it. The isolated location in Maine was in a bowl valley surrounded by mountains which was also chosen with a view to minimize the risk of terrestrial microwave interference.
How did you go about solving these challenges?
You are really testing my memory. Success is easy to remember. All the steps that led to it are not so easy to remember. There were crises. I do not remember that there were any big problems on the fundamental level. Reliability was a constant background concern. We were thinking commercially. The first satellite was never pretended to be anything but experimental, and we were delighted that it lasted six months. We would have like it to have lasted longer of course, but I think we would have been pleased if it had lasted a month. We had no high expectation of long life, but would have been very dismayed if it had failed on launch or something of that sort. We tested every component. We shook them and racked them in every way imaginable and stressed them electrically and physically until hell wouldn’t have it to satisfy ourselves that it was going to be rugged enough to withstand the vibration of a rocket launch and survive in outer space.
The space environment concerned us too. We had no experience operating at close to absolute zero in a high vacuum. That led us, for example, to put all the electronics into a hermetically sealed encapsulated container. That was not the greatest solution in the world, but at least it wasn’t in a vacuum. Then we put sun shields and vanes on it so that the sunlight and the adjustable vanes would keep the electronics at a reasonably constant and comfortable temperature. Periodically in the course of a satellite’s rotation it is going to pass in the shadow of the Earth. Then there is no sunlight to warm it up. Therefore it had to be able to withstand a tremendous range of temperatures. Again the answer was to encapsulate it in a hermetically sealed compartment and get as much sun warmth to it as we could.
Another problem was that even though we had extremely sensitive receiving antennas on the ground, the satellite did not have much transmitting power. Power was derived from a solar cell array. Just a few watts were all the power we had, and we had a traveling wave tube to transmit the signal to Earth. The satellite was spin-stabilized. It was not attitude-oriented. The satellite transmission was essentially omni-directional, so we had no antenna gain there. We wanted to get as much transmitter power as we could, because we had no antenna gain. The solution was to charge up nickel cadmium batteries during the time the solar cell array was in the sunlight. Then the batteries were drained for the relatively shorts period of time – twenty-five minutes or a half hour of transmission – during those limited periods that the satellite was mutually visible between Europe and America. It didn’t operate on a full-time basis, but just in those intervals.
Long-distance transmission systems; digital transmission systems
You moved on to toll transmission work after Telstar.
Yes. Perhaps as a result of the success of the TASI terminals and Telstar I got another promotion. At that point I was made executive director of Bell Laboratories in charge of all the long-distance transmission systems. The very earliest optical fiber development was not started until somewhat later, but we did have all the microwave radio and two or three generations of coaxial cable development. Perhaps even more importantly, my division in Bell Laboratories became responsible for the first terminals for short haul digital transmission in metropolitan areas. These were the first systems that were really digital transmission systems.
All the processing in the TASI terminals was digital, but the transmission over the submarine cables was still old-fashioned analog. Once a speaker had been connected to that cable, talking across it was done the way people had talked across coaxial and past cable systems since the birth of telephony. In the late fifties and early sixties while I was engaged in the speech interpolation and Telstar developments, in a laboratory near my own at Bell Labs, there was the first development of metropolitan digital transmission systems where transmission over the copper pairs was digital. It was at 64 kilobits per second per voice channel. I was not part of the original development of that. Fred Andrews was involved with that and Eric Sumner was the department head. They developed it up to an experimental point.
After TASI and Telstar I was promoted and transferred to Merrimack Valley and put in charge of all the long distance domestic transmission systems. I was also responsible for the production version of the metropolitan digital transmission system. It was the most successful development with which I was ever associated. We made digital metropolitan terminals for millions of channels through two or three generations of development. That’s a great satisfaction if one’s interest is in manufacturing and development as well as experimental research. There were times when the Merrimack Western Electric factory was working around the clock producing digital metropolitan terminals. That was a very successful commercial development and very gratifying. We improved the terminals and reduced costs through a series of developments over the years, all for the purpose of transmitting twenty-four channels – and later forty-eight channels – over copper wire pairs in a metropolitan environment.
Technical challenges on Telstar project
Let’s talk some more about the technical challenges faced in the Telstar project.
There was one challenge with Telstar I forgot to mention. The ground antenna was a horn reflector of 60-foot aperture that transmitted at 6 gigahertz and received at 4 gigahertz. That gave it an extremely narrow beam. The half-power beam width was a very, very narrow and we had to direct that at a moving target. It was a little bit like a skeet shoot. We had a needle-sharp antenna beam and a satellite moving across the sky at a fairly good pace. A major problem was keeping that very narrow beam on the satellite. The first problem was picking it up, getting ourselves onto it and staying on it once we got it. We did a belt and suspender job on that problem.
First, had a computer-predicted track. We had computer-generated predicted altitude and azimuth pointing for the antenna. But we were very skeptical of the adequacy of that. To find the satellite in the first place we used a much lower frequency tracker. We had a quad helix antenna that would pick up a much lower frequency beacon. The satellite carried a beacon transmitting something like 100 megahertz. I have forgotten the exact frequency, but it was not in the gigahertz range at all. The steerable quad helix antenna could pick that up, locate the satellite and lock on to that beacon. Then we could slave the big antenna to the quad helix antenna as a means of tracking.
We had also built a multimode auto-track system into the big antenna itself. We thus had a computer-predicted path to which we could direct the antenna and we could slave it to an auxiliary antenna tracking the low frequency beacon. As it turned out they all worked, but we were very leery. That was the reason for the multiple approach. We did not want to find ourselves groping around for the satellite and then getting onto it and losing it every couple of minutes. That would not have been a very impressive demonstration. Tracking the satellite was a big challenge. Hence the belt and suspenders approach. Each system worked individually and they all worked collectively very well.
Position as executive director; development of coaxial systems, microwave radio, over-moded millimeter waveguide, satellites, wire pair local systems
That’s an interesting approach. It sounds fascinating. Would you talk a little bit more about your work as executive director at Bell Labs?
At Merrimack Valley we developed two successive generations of coaxial systems. We developed the L-4 and finally we developed the L-5 system. By the time we did the L-5 coaxial system we could transmit a 60-megahertz bandwidth and 3,600 telephone channels. They were both technically successful developments but never saw tremendous commercial application. They were installed, and we built systems from New York to Miami and to a number of other places, but they never rivaled the total transmission capacity or extensive deployment of the microwave radio systems. I had responsibility for the microwave radio developments too so I had that as a fallback satisfaction.
We developed successive 4- and 6-gigahertz radio systems, constantly increasing the capacity far beyond that of the original systems. My memory does not allow me to quote exact numbers, but we pushed the developments on and on. One of the culminations of the microwave radio relay developments was the single sideband system. We developed that in the late 1960s and early 1970s.
The quest for reliability was in the background of all these developments. The Bell System was often criticized for slowness in doing certain things, and that rankles veterans who worked there. A lot of the reason for the apparent slowness was the determination that we were not going to put anything of low quality out in the field. It had to work. The reputation of AT&T was largely built on its reliability. We had to have reliability of a different order than most electronic developments of that time.
The transmitting power in the early microwave systems was not very high. It was on the order of about one-half watt. However it had to be pushed as high as possible so that the relay towers could be spaced as far apart as possible on the line of sight while still achieving and maintaining a good amplifiable signal to noise at the relay receivers.
This meant pushing the transmitters into a decidedly non-linear region.
The signal levels of coaxial cables are very low, with repeaters every mile or two. With a thousand of them in tandem the signal had to be kept very, very low in order to avoid excessive non-linear intermodulation distortion. We were unable to do that with the radio systems because we could not space the repeater stations so close together. They were typically 25 miles apart. Therefore the transmitters were pushed well into the non-linear region. In the very first system of the late 1940s the decision was made that frequency modulation rather than amplitude modulation was the appropriate approach. With amplitude modulation and transmitters pushed into the non-linear region the intermodulation noise in multi-channel telephony would have been completely intolerable. It was therefore assumed right from the start that the microwave radio systems would be frequency modulations, FM systems.
As we pushed the capacity higher and higher the FM modulation index had to be smaller and smaller. Transmission began to approach – but not reach – AM type transmission. FM is great for combating non-linear modulation when it’s wide deviation FM but not when it’s narrow deviation FM. Some people began to reexamine the problem, a colleague at Merrimack, Adolph Giger, in particular reexamined the assumptions behind the single sideband AM versus FM systems. His conclusion was that perhaps we could do an AM single sideband system. The lure was that this would increase the capacity of the existing microwave routes enormously. If only 4 kilohertz per telephone channel was required the assigned microwave radio bands could be used to get much greater capacity. I must say as a footnote here that this does not even approach the capacity of optical fibers. We were not in that league at all. But optical fibers were not in sight at the time.
We reexamined the problem and came up with two ideas. Feedback had been used for decades to reduce distortion and linearize amplifiers. Coaxial cable systems would have been completely impractical without negative feedback amplifiers. We could not provide feedback around microwave repeaters or microwave receivers because of the frequency but a fellow named Gene Gordon came up with the idea for using feed-forward. Feed-forward to reduce distortion is an idea as old as feedback. Harold Black, the man who first came up with the feedback idea for distortion reduction also had the feed-forward idea, which was to cancel out distortion by a side path that amplified and subtracted out the distortion in an amplifier.
Gordon built and tested a microwave feed-forward transmitter, and it worked. A very big reduction in the distortion of a traveling wave tube was achieved by using feed-forward, so we started development along those lines. It worked, but it was very, very touchy. The adjustments were very delicate and the slightest imbalance would cause a loss of the advantage of the feed-forward. Our conclusion was that we could get a big improvement in linearity by feed-forward on microwave systems but not as much as we needed.
Then the attention shifted to figuring out how we could reduce the amount of distortion reduction required. We started looking at lower distortion traveling wave tubes, lower distortion in up and down converters, and things of that kind. As it turned out there was a great deal of ground to be gained there. We succeeded in making the components more and more linear. There is a great advantage to revisiting old assumptions in this sort of thing. Finally we got things down to a point where the distortion was almost, but not quite, good. Good enough for AM transmission. Then someone came up with another idea: a solid-state pre-distorter to cancel out the inevitable distortion that was left even when we used the most linear components available. Lo and behold, we got down to the point where we could go back to AM single sideband transmission on microwave radio without the touchy feed forward.
We built, designed and deployed an AM single side band microwave system and it indeed had much greater capacity. I cannot quote you the numbers, but it had much greater capacity than the previous FM systems. Now we are talking about developments in the early 1970s, however, and other things were beginning to loom that made it look as if microwave radio would be supplanted. And eventually it was largely replaced with optical fibers.
A development that I neglected to include in my résumé, but should have, is the over-moded millimeter waveguide. I spent a lot of time on that. After the success of the metropolitan digital transmission systems – in which I was deeply involved with the large-scale manufacture of the later generations – people began to think, “Gee. The world was going digital.” There was the development of digital time division switching arrangements at this time, but the only transmission system we had was the metropolitan one. It was good up to perhaps 40 or 50 miles, but not good for longer distances or thousands of miles.
Rectangular cross section dominant mode waveguides were commonly used on radar during World War II. However there was another wave guide development with a long history at Bell Laboratories. A man named George Southworth that worked in the research department had discovered that in a waveguide of circular cross section in the circular electric mode of transmission, the loss – unlike that in every other medium including coaxial cables and copper wire paths – decreased with increasing frequency. The circular electric mode would be generated at higher frequencies than the transverse field modes. In a two inch diameter cylindrical waveguide frequencies from about 30 gigahertz to 100 gigahertz were used. These were very high standards of electronic transmission systems though not by optical standards.
Southworth had worked on this for decades starting in the 1980s or even earlier.
In the post-war years there had been an early development. It had not gotten very far though it was a very active research project. The problem, as I recall, was the reliability or lack of it, with "hot", i.e., vacuum tube electronics. With the advance in high frequency solid state components there was a renewed effort to bring this approach to a commercial level. I inherited that job with William Waters as director. The impetus was to develop a long haul system of very high capacity, and preferably digital in its transmission mode so it could interface with the local digital systems and the digital switches that were beginning to pop up everywhere. The system was called the millimeter circular electric mode waveguide. We poured a lot of effort into that over a long period of time and spent a lot of money on that project. And it was perfected.
Well, perfected may be too strong a word. It was technically successful. We built several kilometers of guide here in New Jersey and demonstrated successful transmission at frequencies in the 30- to 100-gigahertz range and it was digital transmission. However the development was completely overtaken by optical fibers and rendered obsolete by contrast. It had tremendous technical problems. Going around curves or sharp turns in circular electric mode guide was a very delicate business. That was a problem. Optical fibers had no problem with that.
Were there any other projects with which you were associated?
I ended up back in New Jersey and for a while I was again in charge of satellite development at Bell Laboratories. We had a group working on satellite systems in Bell Laboratories, but all of the actual construction and work on the satellite itself was contracted out. Following Telstar, we never attempted to build a commercial type satellite or ground station. Those jobs were contracted out to Hughes and other outfits.
I was also involved with local transmission such as voice frequency loops and things of that sort. A persistent and nagging problem in telephone transmission is expressed by the famous expression “the last mile.” We got the switching and the long-distance transmission systems thoroughly linked and with optical fibers, satellites and microwave radio transmission was very efficient and economical, but the last mile still was a pair of copper wires to the vast majority of houses, even to this day. Nowadays some people are beginning to have cable access.
I worked for a while on the wire pair local systems. Then in 1978 I was reassigned to a small staff group. The first rumblings of AT&T’s dispersion and divestiture began to appear on the horizon and Bell Laboratories was required to do much closer accounting of its projects in their infancy. For decades we had been pretty much left alone by our owners, AT&T and Western Electric. We called the turn on technical developments and developed the systems we thought were appropriate and they were glad to get them. As the new environment developed more and more competition appeared as well. I was in a project planning group with half a dozen or so people. I didn’t care for that job at all. It was a administrative type job with lofty connections. For a while I reported to the president of Bell Laboratories, but it was not an inline technical job leading to manufacture and use.
You enjoyed technical more than executive work.
Very much so. To a certain type of person – myself included – there is no greater satisfaction than engaging in a development and then realizing that a big factory is working 24 hours a day to turn out the product that your group designed and developed and making money doing so.
That makes a lot of sense.
I think I’ve covered the ground.
Assessment of career achievements
There are some general questions I’d like to ask.
You mentioned that you felt your best project was the TASI project.
Yes. In many ways it was the most satisfying.
Looking at the entire span of your career, what project would you say gives you the most pride? What technical accomplishment means the most to you?
Pride is a funny word. I guess I would avoid using it. I am proud of the Telstar satellite, but it really was a hairy business. There were many aspects I have not mentioned. NASA launched its own communication satellite development and there was pell-mell competition to see who would get done first. I didn’t care for that. I don’t think that sort of thing leads to good engineering. However the satellite was successful and I was in charge of it and I take pride in that. It certainly is the thing that made me a Fellow of the IEEE and got me awards and honorary doctorates from two or three colleges and institutions. It also introduced me to friends that I have to this day who are in England, France, Germany, Italy and Japan. It had a lot of satisfactions associated with it.
From an engineering point of view, the TASI job is the one I liked best of all. The group was of reasonable size with twenty-five or twenty-eight people and I came to know all of them very well. One could be involved and understand every aspect of it, and really I like that. On the Telstar project there were hundreds and hundreds of people involved. I’m not exaggerating to say that I did not know what half of them were doing half the time. I knew, of course, what they were trying to do, but there wasn’t even time to review. This is one of the reasons I dislike pell-mell development. There wasn’t time to review where the satellite tracking fellows were going, what the low-noise receiver group was up to and so on. I could not keep track of all of it. I delegated to highly competent people and prayed that they knew what they were doing. And they did. I can’t take much credit for that. I literally was not involved. I said, “You are the guy, that’s the job, we understand each other” and then he was on his own.
It wasn’t that way with TASI. With TASI I knew at all times what was going on in speech detection, digital memories, signaling and coordination, and in the digital processing switch. There was great satisfaction in that. In terms of satisfaction, TASI gave me the most.
That is probably a better word than pride.
Right. It didn’t get me as many awards or as much recognition but it was more satisfying. I would say a close second to that was watching the metropolitan digital terminals go into really large-scale production. There was a lot of satisfaction in that.
Was this T1?
It was. D1 were the terminals and T1 was the whole system.
Assessment of communications engineering developments
Okay. Looking back over your forty-two-year career what do you think were the most significant developments in the state of the art in communications engineering?
I would have to put the appearance and progression of optical fibers very high on that list. There were groups working on optical fibers when I retired, but they had just been at it for five or ten years. I was angry with them. I was very sore at them – not at the time, but in retrospect I became unhappy. We were working on the over-moded millimeter waveguide and had put a great deal of effort and money into it and it was completely displaced technically and practically by optical fibers. Why the hell didn’t they develop a practical fiber ten years earlier and spare us all that futile effort? I know the answer, of course. I knew those fellows and they were working as best they could. But the realization of a very low loss fiber was by no means certain for a long time.
Another thing I feel has a very high significance is the growing digitalization. We hear that word all the time, “This is digital,” and people saying it do not even know what the word means. It’s just a catchword nowadays. Time division digital switches, the digital short haul systems followed by the optical fiber systems which are all digital in transmission, that’s a major thing. It permits data, video, voice and any kind of signal to be intermixed without any concern. They are all digits. That’s a big one.
I think satellites are a very significant development, but I don’t think they have had the revolutionary impact that these other developments have had.
Collaborations with engineers
Would you talk about some of the engineers with whom you worked that stood out as topnotch?
Near the top of that list I have to put John Pierce. Arthur Clark is credited, and rightly, with the first proposal for a geosynchronous satellite. And Clark is very good and very bright. He is still around. He did it more or less as an exercise of imagination. It was constructive, but the necessary components were not available and he didn’t have a way of getting power in the satellite. Clark said, “If you can get a satellite up to 23,000 kilometers” – or miles or whatever – “it will be in a geostationary orbit.” Ten years later Pierce said, “We really have the makings of a real communication satellite system. We have solar cells for power, solid state for reliable economics and traveling wave tubes with the required bandwidth and power for a transmitter and ultra low noise receivers." Clark had the vision and Pierce saw its practicality. Pierce pointed out that rather than just being forward-looking visionary Clark’s concept was probably practical. He did that in 1955 before we had launched any satellites. He ranks very high on my list of outstanding people.
Close to him I would put Rudy Kompfner, the inventor of the traveling wave tube. I always think of these fellows as paired.
I worked for many years with a man named Charles “Chuck” Elmendorf. He’s a great guy and is also still around. He had two remarkable qualities. One was that he could go into a very disorderly situation and reduce it to systematic order. He worked on coaxial cables and several other projects. After I had been feeling, “Oh my God, this is a ferocious mess,” he would organize and point out the orderly path. Then I thought, “Why didn’t I think of that?” He was also the first fellow I knew that looked at complete systems and not just pieces of building blocks that are patched together. He carefully budgeted the requirements whole systems and put them over.
Another fellow with a lot of influence on me was A. C. Dickieson. He died just a year ago at ninety-five years of age. He had a completely different personality and style than Elmendorf. Elmendorf was systematic and Dickieson was a gambler. He would say, “Damn the torpedoes. Full speed ahead.” He was the first one in charge of satellites before me and gave the job to me. I was more than scared silly of the satellite job because there were so many ways it could fail. I am not exaggerating. Dickieson was bold and would say, “By gosh I’ll do this.”
Going way back to an even earlier inspiration was a man named Harald Friis. He was a radio pioneer, a great engineer and a great human being. He had a tremendous influence and inspired a whole generation of radio engineers and radio designers.
Speaking of more immediate associates, I was very close to Irwin Welber. Welber became a vice president of Bell Laboratories and later president of the Sandia National Laboratories. He’s still around. He is going to call me tomorrow and remind me that it’s the anniversary of the satellite launch. Irv Welber is a great engineer, a great administrator and a great guy in every respect.
Another one was Hugh Kelly. He was a “can do” guy of the first magnitude. He had a big part in the development and realization of the satellite ground station in Andover, Maine. He then went to France and oversaw Bell Laboratories work there. The French built a copy of the AT&T station in Pleumeur Bodou in Brittany and Kelly was our delegate over there. For his efforts in that regard he got a French decoration and was no doubt kissed on both cheeks when he got it. He was a fantastic, great guy and died several years ago.
Your question is almost unanswerable because of the long succession of wonderful guys and competent engineers with which I had the good luck to be associated. I’ll probably offend someone by not mentioning him or her at this point. I have mentioned only a small fraction of the people I have admired.
I guess it is hard to pick a few people out of such a great pool of talent.
When I wrote my tedious history I wrestled with that problem. I decided that since these were group efforts it would be misleading to identify only a few people and by implication slight the many others who were involved. I think I carried it too far.
IEEE, Communications Society
Would you talk about your involvement with the IEEE and the Communications Society?
To tell you the truth, I have never been terribly involved with it, though I have been in it and active at various times. I joined it rather late. My excuse, if it is an excuse, is that I was always too busy to be active in the IEEE. I had a very busy career and I had very little time for my family, being married with four kids. I took very little time for things other than my job. I don’t consider myself a workaholic by any means, but I did take tremendous amounts of paper home. My evenings were often spent – especially as I got executive responsibility for things – trying rather vainly to keep up with what was going on by reading and trying to bone up on things. I got into IEEE pretty doggone late. An associate of mine came around while I was at Merrimack Valley. I had been a student member and then got rather careless about things. This fellow said, “I’d like to nominate you for Fellow grade in the IEEE.” I said, “Oh, that sounds very nice.” He said, “But first you have to join.”
I did join, and I am a Fellow of the IEEE and am very happy with that recognition. I know a lot of the people who are active. I’m close to Amos Joel and Fred Andrews, and I knew Eric Sumner who became president of the Communications Society. He wasn’t president of IEEE or was he?
I think he may have been president of the IEEE also, but I’m not sure.
Anyway, I knew these fellows very well. For a while I sat on an advisory committee to the Board of Governors back in the late sixties and early seventies. I enjoyed that involvement. We didn’t meet too frequently, and we discussed and wrestled with interesting questions and the future of the profession and where it was going.
Future of the communications field
By way of wrapping up, would you talk about your view of the future of communications and communications engineering?
The teaching of it?
The future. Perhaps some of the technical challenges engineers will face in the 21st century.
I should hesitate to say much about that, because I don’t really have much in the way of perspectives. Reading the newspapers we read about the current woes of Lucent and others. I have very strong feelings just as a human being – not as an informed observer – that this is a temporary state of affairs. This is my indirect way of saying, “You ain’t seen nothing yet.” With cellular wireless technology, the capacity of optical fibers and the prospects for increase in that capacity and so on. I’m sure the final overcoming of the handicap of that last low-capacity mile to bring really high capacity into people’s houses will occur. I don’t think we have seen anything compared to what the future is likely to bring. I think information in unimaginable quantities will pour all over the globe and be brought into people’s homes at their command. I’m not sure I regard that with unalloyed joy, but I probably won’t live long enough to see a great deal of it.
I started doing email about five years ago on a very elementary system. A couple years ago my kids, who are all computer savvy, bought and installed for me a little laptop computer and printer. I have four children and two daughters-in-law, and it’s a rare day that I don’t have email from two or three of them as well as with two or three old associates. My routine is to get up, get myself washed up and shaved, prepare a little breakfast, and the next stop is the computer. I would judge that I spend at least an hour on it on an average day. I have to ask myself if this is the best way to spend my time.
What are we going to do when all this information is brought into your desk and your terminal? I can imagine people getting incredibly sedentary sitting there while the world is brought to their door. There is something satisfying about getting out of that chair and going someplace.
The answer to your question in short is that I think it is really almost unlimited. I foresee a tremendous expansion coming. We are in a lull right now, but it’s temporary.
You are optimistic about the future.
That’s a good final note, isn’t it?
That’s great. Thank you very much.
You are more than welcome.