Oral-History:Edwin Harder
About Edwin L. Harder
Edwin L. Harder was a power engineer, inventor, builder of the Anacom computer, and former president of the American Federation of Information Processing Societies. Harder addressed many challenging and interesting problems of the 1930s, 1940s, and 1950s — especially in the regulation and control of power systems. Some of these problems involved the need for additional control as power grids were interconnected and increased in scale; others concerned the regulation of new power machinery (generators and motors) and systems incorporating them (steel mills, paper mills, and so on). His most important contribution was to the Anacom computer. Although he did not conceive the machine, he made such fundamental contributions to the design and construction of the full-scale version and to the management of its operation for twenty years, there is no question that he was the person who made computing an important activity at Westinghouse.
About the Interview
EDWIN L. HARDER: An interview conducted by William Aspray, IEEE History Center, 30-31 July 1991
Interview # 118 for the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc.
Copyright Statement
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, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. 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:
Edwin L. Harder, an oral history conducted in 1991 by William Aspray, IEEE History Center, Piscataway, NJ, USA.
Interview
Interviewee: Edwin L. Harder
Interviewer: William Aspray
Place: Dr. Harder's home near Pittsburgh, Pennsylvania
Date: 30-31 July 1991
Family Background and Early Life
Aspray:
This is an interview on the 30th of July 1991. The interviewer is William Aspray of the IEEE History Center. The interview is with Dr. Edwin L. Harder in his home near Pittsburgh, Pennsylvania. In this session we'll talk about the early life and career of Dr. Harder. Could you tell me when and where you were born?
Harder:
I was born in Buffalo, New York, on April 28, 1905.
Aspray:
And were you the only child or were there other children in your family?
Harder:
I have an older sister, and then I have two younger sisters and a younger brother. There were five of us, four of whom are still living.
Aspray:
What did your parents do?
Harder:
My father was a superintendent of the stations of the Cataract Power and Conduit Company in Buffalo. He had graduated from Penn State in electrical engineering. He was an electrical engineer and had worked in Philadelphia and New York briefly, and then he came to Buffalo at the time of the Pan-American Exposition in 1901 with his new bride, my mother. They lived on 212 Virginia Street, which I just found out a few weeks ago, and built a house at 28 California Street where I moved when I was two or three months old. He had his offices in Terminal House B where the lines from Niagara Falls came across the Niagara River to the terminal house.
Aspray:
Were there electrical engineers at the time your father became one?
Harder:
No. The custom was pretty much to hire men off the street to perform all these functions. So my father had a telephone beside his bed, and he would have to get up anytime of the night to go out if there was a manhole explosion or some other malfunction and explain to these workmen how to fix it because they didn't know. College graduates were very rare in those days. In 1901 just six years out of college he went to Buffalo to head up the substations. Paul Lincoln did the same thing at Niagara Falls. Westinghouse sent him to Niagara Falls when he was only a few years out of Ohio State to be in charge of the largest power plant in the world. Only the college graduates knew how it worked, and my father was one of them.
Aspray:
What were your living arrangements at that time?
Harder:
Well, the house on California Street was a two-family house. We occupied the downstairs and rented out the upstairs. Went to the nearby public school, about three miles out from downtown Buffalo. The Italian population was gradually expanding north, so eventually we moved, and that area became mostly Italian. We moved a few blocks farther north, and I went to Lafayette High School in Buffalo. Then from there to I went Cornell University.
Aspray:
Was Lafayette a good high school?
Harder:
Yes. It was one of the best. It was good for college entrance. There was a technical high school in Buffalo, but Lafayette was a scholastic high school, and it taught all of the things that were needed to enter college. A good many of its graduates went to college.
Aspray:
What were your parents' expectations for you?
Harder:
Well, my mother believed in keeping me busy. She always found jobs for me. I worked in a west-side fruit and market company after school and Saturdays to keep me busy. I think my mother thought it was better for a boy to be busy.
Aspray:
Were there expectations about what kind of career you'd have?
Harder:
No, I don't think so. I think I just naturally followed my father's career. Through the third year at Cornell I could have become a mechanical engineer. I had given it serious thought. So it wasn't until the fourth year that I had to definitely decide to be an electrical engineer.
Aspray:
Did you have any hobbies when you were young?
Harder:
In Buffalo in those days, of course, many of the kids had roller skates and scooters. In the winter Buffalo freezes, and it was all horses and wagons, and they never cleaned the streets, so there was ice-skating all over. In the parks, in the big circles, they would bring the fire hoses and just freeze them over for the kids to ice skate on. So bicycling and roller-skating and sled riding and ice-skating I remember particularly as hobbies.
Aspray:
Did you have interests in science when you were a boy?
Harder:
I was a ham radio enthusiast, and so I heard Pittsburgh on the set in Buffalo before I ever moved to Pittsburgh. That was the very early KDKA broadcasts.
Aspray:
You built your own equipment?
Harder:
We built our own equipment. Earphones were kind of expensive. But my pal had a friend who was a telephone repairman, and he had an earphone that he would loan us occasionally. And later on I acquired a set of earphones, but this boy and I were close together in ham radio and many other things and then became roommates in college.
Aspray:
Did your father take part in any of these activities with you?
Harder:
No, he came home late from the office and didn't enter into our activities very much. He was interested more in charitable activities. They had a Prisoners' Aid Society in Buffalo, which he practically ran.
Aspray:
Did you get a chance to go down to the power company and see his workplace and learn about what he did when you were a child?
Harder:
The last year before I went to college I decided I really ought to work in the electric company. So I worked in the meter department of the Buffalo General Electric. I should say that the Cataract Power and Conduit Company was 25 cycles. Twenty-five cycles is better for transmission. At the same time in Buffalo there was a Buffalo General Electric Company that was 60 cycles. Eventually they merged, and so my father was in the Buffalo General Electric Company from then on. And all of the city had to be changed to 60 cycles. They went through it block by block, gave everybody 60-cycle appliances to replace the 25 cycle ones and changed it all to 60 cycles. [My father did enter into certain things at certain times of the year. I remember on Memorial Day I was always wanting to go to an amusement park or go boating or swimming. But we always had to first go and pay our respects to the Civil War veterans that were still marching in the big parade down Main Street because my grandfather had been a drummer boy in the Civil War. When I was ten years old, those who were not over 20 in the Civil War were just 70 then. They were still marching. So we had to go down and see the Civil War veterans march, and my father would take us. Also to Crystal Beach, which was the amusement park near Buffalo. Crystal Beach was a 25-mile lake ride out to the Canadian shore the beach and back. The steamship ride was a big part of the amusement park adventure. So my father did enter into activities to some extent. And he liked these activities. He was just so busy with other things he didn't spend a lot of time with us.]
Aspray:
When you were growing up at home, was there an expectation on your parents' part that you'd go off to college?
Harder:
I don't think that was ever mentioned that I recall. I just assumed that I was going to college when I finished high school. My father would have liked me to go to Penn State, but the State of New York gave many tuition scholarships for Cornell. There were eight given in Erie County alone. I got the first one, and that paid the full tuition at Cornell for four years. So that settled the question of where I would go because so much of the finance was covered. I did very well in studies in school and I was generally at the top of the class.
Aspray:
What was your mother's education?
Harder:
She grew up on a farm up in the Adirondacks. She went to teachers' training college nearby. She also had some nurses' training. She must have had some secretarial training, too, because she was in New York and working for a firm that supplied part-time secretaries. I know at one time she went as a secretary up at GE while some of their help was on vacation. So she was quite well educated and taught school some. She met my father in New York when he was burned in a manhole fire. She was in the hospital at that time, and they got acquainted.
Cornell
Aspray:
Tell me about your course of study at Cornell.
Harder:
Well, it's pretty much the same as electrical engineering today except that at that time it was not assumed that all the physical activities were known in industry. So we took foundry to make cast-iron. We took forge to turn it into something. We took wood-working to make the patterns for it. And we took a lot of mechanical drawing. These were courses that didn't require any homework. You'd go and do them. They were physical things that you do. And nowadays it's assumed that all that technology available out in industry. You don't have to teach it in college. But they were still teaching it at Cornell when I went there. Then there were two or three years of very strong physics and chemistry, and that was very valuable to me in the later patents and work that I got into. When we finally got into electrical, we had Valdimer Karapetoff who was the principal professor and certainly one of the best known in the country and one of the most capable. I was amazed at the ease with which he could do things. I think it was a great boon that I studied under Karapetoff at Cornell. I think that helped me later.
Aspray:
What kinds of training in mathematics did you have?
Harder:
In college, I worked up through differential equations, starting with algebra. I remember the math professor when I was a freshman apparently knew all of the math teachers in the high schools all over New York State. And he said, "Where'd you come from?" I said, "Lafayette High School." And he said, "Oh, Hallie Poole, eh?" [Laughter] He knew them all.
When I came to Westinghouse, I probably had a better grounding in differential equations than most students. I learned this when one of our instructors from the General Engineering Department, gave us a problem: Derive the equations of traveling waves on transmission lines. Well, I scratched it right off in differential equations. And it surprised him. He marked it "Very good!" I don't think anybody else in the class could have done it. And I think that's why I got into that department eventually, because he knew that I could handle the kind of work that they were interested in. So the math at Cornell would go up through integral and differential calculus — and differential equations. Then Karapetoff taught what would be called operational calculus. A fellow by the name of Berg has written a book on operational calculus. I think now you call it Heaviside's operational calculus. It's sort of cut-and-dried calculus. There are certain algorithms — that are used, and Berg and Heaviside apparently found — them and used them in England, too. So I had a smattering of that before I came to Westinghouse. Then they also taught a little bit of symmetrical components.
Wagner and Evans and Fortesque at Westinghouse were writing papers on it at the time I was at Cornell. And we had some of those papers to study.
Aspray:
How did you choose electrical engineering as opposed to mechanical engineering or physics?
Harder:
Well, I think by that time I just liked electrical better than mechanical problems. I knew pretty much from the time I went off to college that I was moving in an engineering direction.
Aspray:
Tell me about your courses in electrical engineering.
Harder:
In the second and third years you get electricity in physics. You don't know that there is such a thing as a three-phase circuit. Circuits are all single-phase in physics.
Aspray:
Right.
Harder:
In Buffalo many of the engineers were men that came off the street, but they knew that circuits were three-phase. And here I came home from college about the third year, and I still didn't know that circuits were three-phase. They thought I was pretty stupid coming from Cornell in electrical engineering and I didn't know that circuits were three-phase. Then in electrical engineering at first you study DC circuits and resistors. Then you study alternating-current circuits. But still the fact that a lot of them are three-phase or polyphase circuits and they don’t enter in at first because you have a lot to learn about network solutions. There are various ways of solving them. The only ways I learned in college were by equations. You write the mesh equations for each mesh. And if you have ten meshes, you get ten equations, and you can solve them. When I came to Westinghouse I learned that there's a lot of other ways of doing it, too, we used the equation that until 25 years later when computers became available. Then we went back to the inverting of big matrices as a way of solving equations. But in college it was DC circuits and then AC circuits, and all kinds of problems.
Karapetoff had a system of teaching which was good. The lectures were large, and then a week later, after the lecture, you had a recitation with an instructor and only ten or twelve people in the class. You got right down to business with them and worked out problems on what Karapetoff had taught in the lecture. Then the next week you had it in laboratory. So he hit you with the same idea three times. His theory was that if you got hit three times with it and didn't understand it by then, you'd never understand it. Of course at the beginning of the semester it was a few weeks getting into this rhythm, and at the end you had to work out of it some way or another. All through the semester we had that rhythm of getting the same subject in the lecture, then in recitation, and then in laboratory course. The laboratory experiments were time-consuming. They were quite long reports, but not so much, though, as what they called "mech lab," mechanical laboratory reports all during the second and third years. They were the reasons that Cornell gave degrees in M.E./E.E. instead of just a B.S. degree in science because they were equivalent to a thesis. They were spread out — about twenty hours per report, but one every week. All long experiments in mechanical laboratory and then later lesser ones during the electrical career.
They had the ROTC at Cornell, and most electrical engineers went with the Signal Corps at Fort Monmouth. But since I could have been a mechanical engineer, I ended up being in Ordnance. That's how I happened to go to Aberdeen Proving Grounds. In four years of ROTC you get a second lieutenant commission, which you had to work on afterwards to maintain. Well, I was taking graduate courses afterwards; I didn't have time for both. So I gave up the ROTC.
Aspray:
How well did you do in college?
Harder:
Well, I wasted a lot of time unlike my roommate. He'd go back to the room and study, and I'd waste time playing chess. I could waste time more ways. I ended up one year with six reports behind. I stayed up six days in a row — never went to bed — and did those six reports and got them in, and they accepted them all. They didn't have to, but they did. Well, in the end, I was one of two in the class that was elected to all three honor societies — Phi Kappa Phi, Eta Kappa Nu and Tau Beta Pi. So it was just a natural ability rather than working at it. I could have done an awful lot better. I could've learned a lot more, and I could've taken a lot better advantage of college. The same was almost true of the first year at Westinghouse. I didn't take things seriously until about the third year.
Student Courses at Westinghouse
Aspray:
You graduated in 1926 from Cornell. For a student finishing up with an electrical engineering degree at Cornell at that time, what were the job options if one went right into work and didn't continue education?
Harder:
Well, you could go with the telephone company, or with an electrical manufacturer, such as Westinghouse of General Electric. Westinghouse for example had either sales or engineering options. You could go with Allis-Chalmers. In fact, I was offered a job with Allis-Chalmers. Offers were very plentiful in 1926. By 1930 they all dried up. The Depression came, and many students getting out four or five years later couldn't find any jobs at all when the Depression hit. Allis-Chalmers invited two of us out there to spend the whole Easter vacation with them, and neither of us went with them. My logic at the time was that I wanted to go with a bigger company that had more varied things to get into. I said then I might go over to Allis later, but I was going with Westinghouse then. I'm sure there were a lot of other offers, but I don't remember them all.
Aspray:
Did you decide to continue your education at that point or did you go work for — ? A Company?
Harder:
I took about a month's vacation my parents didn't entirely approve of. They'd paid my way all through college; so they thought I ought to be anxious to get a job and start working. But I signed up for Westinghouse on July 12th. At that time all students went through a year of student course. We all got $100 a month to start with. After six months it went up to $110 a month. Then you went up into some department after that.
Aspray:
What was the purpose of student courses.
Harder:
For sales students they got a lot of short assignments all over the company. Engineering students got into engineering departments for maybe as long as two months in a place. There was also a three month engineering school. In this class, principal engineers from the different departments would come and lecture and give problems, for this we received credit at the University of Pittsburgh. I think it was 15 credits. So to get a master's degree, you only needed six more credits. So a great many went on to a master's degree. I got mine in 1931, which was five years after I went with the company. But those men that came and taught this course, their purpose was to tell you about this line of equipment. For instance the fellow on transformers explained how you look at transformers, how you design them and how you work out problems of them. We had very fine engineers from all over the works come for this three months, and they taught us an awful lot.
Aspray:
At the end of that period did you have some say in where you were assigned or were you simply assigned to a department?
Harder:
For the first year you didn't have any say; the Educational Department worked out a schedule. You might have expressed some preference, but I just went where they sent me. I spent sometime in the shop, building switch-boards, wiring. I spent a couple of months setting up for tests for big machines, and a couple of months in motor engineering and design of motors. I went up to Derry for an assignment where they made insulators. The company had taken over a factory that made restaurant china and converted it into an insulator plant. I think even going to Derry was partly due to the fact that I said, "Can't I go someplace besides Pittsburgh?" All my assignments had been in the Pittsburgh area so they sent me up to Derry for a few weeks. I still wasn't taking things very seriously. I was just doing what came naturally and being able to do things without any great effort. My mental ability apparently was there all the time, and so it would come out in spurts. But it wasn't 'til a few years later that I really accepted my responsibilities and became a good engineer.
Aspray:
How large was the entering group of employees?
Harder:
300 from all over the country, about 100 of them engineers.
Aspray:
Were there many from Cornell?
Harder:
Yes. In Wilkinsburg here there's the old Singer Mansion. Anna E. Dixon had forty boarders and roomers there, and there were eight or ten from Cornell. I knew them before I went there, and one of the reasons I came was that some of the ones that I had particularly admired at Cornell had come here. They got me a room up there where they lived in the Singer Mansion. One of my Singer Place friends died a few months ago out in the State of Washington. He was one of my best friends in that class and later built a number of the devices that I invented. I sent him a picture. That's the reason I know there were about 100 in the engineering class because I said, "How many of these can you name without turning it over?" He knew who they all were. But later, during the Depression, there were very few hired. And then during the war they couldn't hire any. They weren't available; they all went in the army.
Well, I had written both to Frank Rhodes, the president of Cornell, and to Street, who was dean of engineering, and expressed to them how important I feel it was to have been associated with some of the top people. I mean, many people can teach the courses, but having Karapetoff and learning the powerful ways he has of looking at things was a great benefit in my later work. Another who came from Cornell was Joe Slepian. He taught math at Cornell, and then he just left and came to Westinghouse and set up shop here. He became one of our very best scientists at a time when nobody knew anything about conduction in gases. And his physics was very fine, and he knew how to study and how to use the things he studied. He lived a couple of blocks from us here in Regent Square. A great many of the early things involving conduction of electricity and gases he invented: lightning arresters, circuit breakers, all sorts of things. He contributed a lot to the work on lightning. In the early days there was not much known about lightning, and lightning destroys a lot of electrical equipment. So we had to know how to design electrical equipment.
One of the great electrical engineers at Westinghouse was Fortescue, who conducted original experiments that led to what is called the direct-stroke theory. Before then people believed that strokes anywhere in the neighborhood could damage equipment. But he essentially proved it was only strokes that actually hit the lines. Fortescue was a great engineer. He contributed a lot to the knowledge of lightning, and also he contributed symmetrical components. He must have been a good mathematician because he understood that these problems that we had of unbalanced, three-phase systems: that you could resolve any set of three unbalanced vectors into three sets of balanced vectors. Out of that grew the whole science of symmetrical components. It takes textbooks to explain it all, so I wouldn't attempt to go into it at all. But we owe that to Fortescue, and he presented it in 1918 as a paper.
Then I left Joe Slepian, who was a strong force in Westinghouse for many years. Some of his devices are in the Science Museum in Munich. He had a good grasp of how electricity went through gases. I took courses from him on conduction in gases. He later became associate director of our research department, but he never became a manager. And there are quite a few great scientists who don't go the road of managers. Fortescue was one. He never became a manager. Slepian sort of directed the technical program. Frank Conrad, who was responsible for the early broadcasting, started in the shop. He became chief engineer at Westinghouse. But whether it was because of that route, he was always an individual worker and inventor. He had 200 patents when he retired. Slepian had about that many, too, but these men never graduated, you might say, from the technical work into managing people. So that makes quite a difference.
Aspray:
Is there more you want to say about the engineering school itself?
Harder:
Only that it had a salutary personal effect in that all through my career with Westinghouse I knew those 100 people, and they were scattered all over the company. So there were people in all the different plants of the company that had been in that engineering school and I'd gotten acquainted with. I never would have if I'd just come to work in one place. I'd have gotten acquainted with one or two, but not as closely. We became good friends as well as co-workers. This fellow in Sharon, Ed Wentz that just died, he built a lot of the things I invented. The invention isn't the whole thing: part of it is building it right, too.
General Engineering Department
Aspray:
At the end of that first year of student courses what happened?
Harder:
I applied to go into General Engineering. There were a lot of design engineering departments, and then there was one general engineering department. And it just seemed to me that fitted what I wanted to do.
Aspray:
What's the fundamental difference between those two departments?
Harder:
The general engineering department doesn't build anything. They work on all the system problems. At the heart of it (at least in electric utility) were nine sponsor engineers. One engineer would be assigned to the Middle Atlantic District, and he would go out one week a month with the salesmen to the customers to help them solve the more difficult technical problems that occurred in the application of electrical equipment to their systems. Then there were the inside men that worked with them to help on the problems. So I was accepted in general engineering. If you went into general engineering, you were required to spend two years first in some design department to learn that particular equipment. I was assigned to the Power Engineering Department, which made large rotating equipment. I went to work in the electrical development section of that department, which was the best place to learn about electrical equipment. Instead I worked on air flow the whole year because the chief of that section was working on the air flow and ventilation of these machines. I didn't learn a thing about rotating machinery. I learned a lot about dimensional analysis in models, which stood me in good stead later when I became manager of the Advanced Systems Engineering and Analytical Department because a lot of people didn't understand that very well at all.
A lot of the technical literature that we had on air flow came from Germany. After World War I the Germans were not allowed to use power planes; they were allowed to use gliders. Consequently they did some very fine model work on gliders and gliding, and we had all this. We were associated with Siemens in Germany, and it came through that route. We got all this literature, and so I learned a lot about air flow. In July of 1928, at the end of the first year there was a general ten percent lay off. Every department had to lay off ten percent. And of course the first thing Power Engineering did was lay off the students — which included me. My boss, however, got me a stay of execution for three months so that I could write reports on all the experimental work I had done during the preceding year. Then he went on vacation. Well, after General Engineering had their ten percent lay-off, one more man left. They therefore had to have somebody right away, so I went to General Engineering after one year instead of two years. When my boss came back from vacation he was furious that I had done this when he'd gone to all the trouble to get me those extra three months to write those reports. I assured him that I could do it. I had no problem at all writing those reports because in college I had written six in six days. My boss was a great guy; it was just a momentary frustration for him.
Later he wrote a paper on the hot-spirit anemometer, which was a little device that I had developed, although it was the invention of one of my predecessors. You could stick it through the bolt hole of a machine and measure the air velocity inside of it. They used a ball bearing for silver shell, and it had thermocouples and heaters. I had developed it. My boss wrote a paper on it and put my name on it as well as his and insisted I take the payment for it. I wasn't married yet, but I was only getting a little over a hundred dollars a month. This was such a nice gesture from an older man that had plenty of money. He was a very nice man.
Aspray:
The system in which you trained of the year of instruction and then in a particular department: how much success did it enjoy?
Harder:
There's a long history of being placed in a department expecting to get training, and then not receiving it. There was an apprentice system for shop people years ago. In Germany and early in the United States if a man went into a shop to work, he worked there for his life, so the company could afford to put quite a lot of training into him.
Westinghouse had this system, and then they trained every man that came in engineering for a year. They didn't get anything out of them in the way of production. But as industry grew and mobility increased, shop people didn't stay in the same shop. They had this training, and they could go someplace else. The same thing happened in engineering and sales. People didn't stay with Westinghouse all their lives. Westinghouse couldn't afford to do this anymore. So the training period became shorter and shorter, and eventually was stopped. Now the whole thing's gone, but I don't think for many, many years they've had a student course. It changed with the times and with the greater mobility of people. At this time Westinghouse and other larger companies invested a lot of time in pure training-a year of student courses for engineers and apprentice courses for shop machines. The unsaid expectation was that a man would stay with that company all his life and the company would receive the benefit of that training. (This does not apply to the subsequent two years for engineers entering general engineering, these were productive.)
Aspray:
What happened when you moved over to the General Engineering Department?
Harder:
I went to work for Bob Evans, who was working on railroad electrification. Railroad electrification's were all 25 cycles. Unlike 60 cycle power systems that could be analyzed using the D.C. calculating board, with sufficient accuracy by long-hand calculations, using desk calculators. The Virginian Railway Electrification, the Norfolk and Western, and The New York, New Haven and Hartford electrification's had all been calculated using these mechanical desk calculators, either hand cranked, or with a motor. Now we were working with the Pennsylvania Railroad, and negotiating with the Reading Railroad. I was involved in those calculations.
At the same time, rectifiers were beginning to be used, and telephone interference had been developing between power companies and communication companies. Before I came with Westinghouse, the Electric Utility Industry and the Bell System had settled a lawsuit for $25 million, that would be $2.5 billion today) out of court. Then they got together and said, "This is foolish paying the lawyers all this money; we've got to correct this thing anyway. Why don't we just get together and determine what's the best engineering solution in each case and do it?" So they formed a triple-joint committee which studied the harmonics produced by rectifiers. I was right in the midst of that. I learned an awful lot about harmonics. Before long I was teaching circuit theory, and I taught a lot about how to analyze for harmonics. I first used a Chubb mechanical harmonic analyzer, like a planimeter that you move mechanically around a cutout of a polar oscillogram. It multiplied by one harmonic after another. so each time around it determined how much of that harmonic there was. Then the electric harmonic analyzer came out. I designed the filters for a lot of the rectifiers, the DC filters and the AC filters, to cut the harmonics to whatever was required for a particular system. I learned a lot about harmonics in that early work. So when working on railroad electrification's we had to calculate them all longhand with these desk calculators because the DC board wasn't accurate enough for 25 cycles.
Calculating Equipment
Aspray:
Can you describe, in terms of the mathematical problem that has to be solved, whether a desk calculator or some sort of analog machine must be used?
Harder:
Okay. You have an electrical network which has a lot of branches with inductance and resistance in each branch. Say, for instance, that the system you're looking at is a generator in Philadelphia that supplies transmission lines, four of them, going towards New York. Every seven miles there are transformers that go down to the trolley voltage of 12,000 volts. Then there are four trolleys for a four-track road. Now if there are 50 meshes in that, you could write 50 simultaneous equations and invert the matrix — solve it. That wasn't the way it was done, however, if you take any one mesh point, you'll find lines, maybe trolleys this way, trolleys that way, transformers up here, and let's say that's a trolley point. That mesh point you can eliminate. You can assume that there's a delta of lines connecting those same other three points, and that delta is equivalent to this star. And so there's a star delta equivalent. So you can solve that network by eliminating one point after another until all you have is a generator and the endpoint that you're interested in.
We developed an alternating current calculating board at Westinghouse, so we didn't have to use desk calculators anymore. The last one that was done using desk calculators was from Philadelphia out to Trenton. It took two weeks with two of us working, steadily, to check each other. We each did the same calculations and checked at each point to make sure we both got the same answers. But we then developed an AC calculating board, and Gibbs & Hill in New York bought one to handle the railroad calculations, and we had one in East Pittsburgh. So for 25 years those numerical methods were not used until digital computers and methods came in and we went back to them again.
Aspray:
Did you build your own equipment?
Harder:
Yes, but the DC calculating board was developed before I came with the company. It was available in the 'twenties. The AC calculating board was built in 1929. Since I had been doing these hand calculations, I could help them by telling them what size branches they should have on this new calculator. You had to have many variable resistors and reactors, and I could tell them what size they had to be in order to do these problems. They built the board in the Switchgear Department. Bill Parker was the engineer who did it, and I just helped him on that, on the AC calculating board. So from then on the AC calculating board was used for all those AC railway electrifications. To go on with the harmonics story: from this joint work with the Telephone Company I'd become familiar with harmonics, and I was teaching them. One day two trains going into New York City were heading for each other on the same tracks in the tunnel. They got stopped, nobody was killed, or hurt. But boy, was there an uproar!
What had happened was that the signals had failed because the Long Island trains came into New York on DC, and the Pennsy trains came into New York on 25 cycles. And the signal power system was 100 cycles. It was assumed that you never get even harmonics on a power system. But with AC and DC trains on the same tracks some of the DC gets into the AC transforms which then produces even harmos 100 cycle with in this case the same as the signal power, As a result the signals failed. The signal power frequency had to be changed. But the operator of the switch and signal was afraid there'd be some intermodulation that would cause that, so he went halfway between this and 100: 91-2/3 cycles. It has worked fine ever since. It solved that problem, but then we had to figure out what to do about all that extra heating in the engines with half load of DC flowing through these things and the current transformers.
Aspray:
Was this the first time you'd ever used calculating equipment?
Harder:
I had never used a desk calculator until I got in General Engineering. They were Marchand machines. They cost about $700 more than an automobile in that day. Several years ago I bought a little calculator out at Sears for $7 that could do a lot more than that $700 machine could do. [Chuckling] I had just used a slide rule in college, which was quite adequate for everything. We never made any important calculations. As a matter of fact, a slide rule was used widely in engineering probably until about 1960.
Aspray:
Were there any analog calculating devices used as part of your education?
Harder:
I'm not aware of any, although I know there were in some universities. In Germany I know they used hydraulic tanks for certain things. We didn't have any at Cornell. I don't recall, but almost any little thing that responds exactly to equations and that you can manipulate has been used at one time or another for an analog computer. Of course, electric analog computers were the easiest to manipulate and the most general of all of the analog computers.
Aspray:
Were those kinds of analogs familiar to you when you went to Westinghouse?
Harder:
I think someplace along the line I became aware that there was a DC calculating board down in the Switchgear Department. They used it for power system calculations, for setting relays and deciding how big the circuit breakers should be and things like that. I have a copy of a 1920 paper the engineers, I'm sure, used little models before that. But there was a formal DC calculating board by 1920 in the company.
Alexander Monteith
Aspray:
Do you remember any of the other people you worked with in General Engineering?
Harder:
Yes. Fortescue and Wagner were consulting engineers, and they had offices nearby, but they were closely associated. As I said, there were these sponsor engineers for the different districts. The fellow that I had most to do [with] in Westinghouse all through my career was Monteith, who had the mid west district that is out of Chicago. Monteith retired finally as senior vice president. Right from the start he was a man that people would listen to. With most of the little power companies out in the Southwest, if Monty told them what to do, that was good enough for them. That's the way those systems were designed. He told them how to do it. Well, Fortescue discovered, by test down in Tennessee, that only direct lightning strokes to transmission lines cause the damage. So Monteith carried it on from there and wrote a paper on line design based on direct strokes. You know how many direct strokes strike the earth and strike a mile of line in a year, you know how big they are. You design the grounding resistance of the towers so that this current goes through over the ground wire without causing such a high voltage that it's going to flash over the line. We calculated it all. After writing the paper Monteith later became manager of that central station department, he got me back in. There had been very early a table of transmission line constants. It was called Nesbit. Nesbit had calculated the impedances of transmission lines for all spacings and all wire sizes that were currently used at that time. It's getting to be time to revise that. But Monteith conceived instead of making a transmission and distribution reference book, and his paper on line design based on direct strokes is one of the chapters of this book. I wrote two of the chapters of it. That book titled Transmission and Distribution Reference Book is widely used all over the world. And so later Monteith became my boss, and a finer one I couldn't have had.
Switchboards and Relays
Aspray:
Tell me about one of your inventions.
Harder:
Many of my inventions were highly technical, or mathematical, in nature. The problem would be expressed mathematically, and I would try to find apparatus that would do what the resulting equations said had to be done. Some of my finest and most valuable inventions were the result of this process. They would be very difficult to explain in layman's language. But not this one that I am going to describe. It was simply a good old-fashioned inventor's invention - a good idea that no one else had thought of - and a lot of good luck. And yet it was worth many millions of dollars to the Westinghouse Company, over a period of several years.
During World War I, the Army Engineers started using carrier current (a few watts of 50 to 150 kc current) to transmit messages over long power transmission lines. In the post-war period power companies took up this technology for communication and protective relaying circuits over their transmission lines. By 1930 both Westinghouse and its principal competitor, in this field, were offering carrier current relaying. When a fault occurred the fault detectors at each end of a line would start transmitting carrier. If the directional relay at terminal A pointed into the line it would stop carrier transmission from that end. If the directional relay at terminal B also pointed into the line, then the fault must be on trip simultaneously, with no delay. Three-phase directional relays were used because the load from the fault current in a faulted phase. In the three-phase relay the fault current, of course, predominated.
The best previous relay scheme used distance measuring relays, but involved sequential tripping for faults near either end of the line, a delay of about a half second for the second breaker. Until 1936 both companies were using conventional three-phase directional elements, operating in 3 to 5 cycles. They were 3-5 cycle carrier systems. The circuit breakers then took about 8 cycles to open. However, unbeknownst to Westinghouse the competitor had spent a full year developing a very fast three-phase directional element and were now offering a one-cycle carrier scheme. It was almost impossible to sell a 3-5 cycle system in competition with a 1-cycle system. Both the Relay Department in Newark, and the Switchboard Department in E. Pittsburgh were in serious trouble. About half of the relays were sold with the switchboards. Many new transmission lines were being built at this time, and Westinghouse stood to lose many millions of dollars worth of relay and switchboard business.
In 1936 I was in the Switchboard Department. General Engineering had been almost eliminated during the hard days of the Great Depression. I had been with the company for 10 years and had made a number of valuable inventions. It was decided to loan me to the Relay Department in Newark for six months or a year to help develop a competitive system. At the first meeting there was a pretty glum bunch of engineers. Not only were we a year behind, but how in the world were we going to develop a high speed three-phase directional relay without infringing their patents? We just had to have some interim scheme to offer in the meantime — no matter how crazy.
I cooked up a scheme using the high speed single-phase directional elements of our distance relays, interconnected with a lot of fault detector elements. On paper it would be one cycle. But eight separate relay elements, four directional, and four fault detectors, all had to open or close correctly, and coordinate with each other, and do it in less than one sixth of a second. The probability of this happening seemed very remote. But no one had a better idea, so we decided to try it out in the laboratory. The laboratory engineer came back in a couple days with the good news that they did seem to coordinate. As a matter of fact in a large number of tests he had not had one single miss-operation. We immediately started to sell that system, and really had the luck of the Irish.
The first customer wanted badly to buy Westinghouse relays, but there were going to be one-cycle carrier relays at the other end of this tie line. He was overjoyed to learn that we could now use the single phase directional elements to control carrier. After that it seemed that each job we sold, the customer would find some new advantage of this system that we had not yet thought of. We were just glad to have something that worked at all-in one cycle. One of the first customers observed that with a three-phase directional element, when you pull the test switch to test the relays the transmission line has no protection. With the single-phase elements, there are four of them-three phases and ground. Any fault always involves at least two (phase-to-phase or phase-to-ground). Thus the test switch can be pulled on any one of the four to test it and still have 100% protection of the line. When all are in service there is some redundancy. Another customer observed that if the carrier was out of service for any reason the scheme reverted to distance relaying, the very best protection before the days of carrier. Still another noted "Well! I can buy distance relays today and add the carrier later when it is justified". And so it went. The 1938 paper lists several other advantages of this new carrier-pilot relay system.
The net result was that at the end of the first year we had sold 87 equipments, and lost 13. This continued for quite a few years. Westinghouse never did develop a high-speed three-phase directional element. The competitors abandoned their system after a few years of dollars worth of business, we gained substantially. Had I not gone to Newark, would the engineers there have come up with this system? That we will never know. But they had been struggling with the problem for several months and had not thought of it. Our competitor did not think of it in the year they were developing the high-speed three-phase element. Relay engineers had long ago determined that a carrier system required a three-phase directional element. A newcomer is uninhibited - has no preconceived notions - in this case with a most fortuitous result.
One of the sequels to that: four or five years before I had invented the HCB relay, which is still in widespread use today, we had a very prominent engineer by the name of Bill Lewis, and a visiting engineer from Australia by the name of Tippett. They had worked together and using symmetrical components, they had worked out the fundamental basis of distance relaying, the most fundamental paper on relaying probably that's ever been written. Well, at that time there were great economies going on in the IEEE and not all papers got to be Transactions papers. Some were conference papers. Unfortunately this very fundamental paper fell to the lot of being a conference paper and was never published in Transactions. So three or four years later when I was chairman of the Relay Subcommittee, I got it resubmitted. I went through the proceedings all over again and got it published and put in Transactions, just because it was such a fundamental paper.
Well, after they had done that work, Bill Lewis put in a patent for a pilot-wire scheme where you use pilot wires between the two stations. He knew exactly what information you had to transmit between the two to handle all faults, and it took four wires to do it. I said, Heck, you don't have to use four wires. You can do that with two wires. What you have to do is at each end you use a positive-sequence network that produces a single-phase voltage, proportional to the positive sequence. It doesn't matter what phases it's on. Fault on any phase has a lot of positive sequence, and to that little resistor voltage you have to add a voltage with a ground current through it because the ground faults are often very much less. They don't ground systems solidly; they ground them through a high impedance sometimes. So it ended up that you can easily produce a single-phase voltage at each end that's proportional to the positive sequence and some constant times zero sequence. All you had to do was compare those two; you don't need four wires. So I immediately put in a patent on this, which I called the HCB. Things were very poorly organized for patents in the early days. I don't think anybody told Newark that such a patent even existed. They didn't sell many pilot-wire relays anyway. They may not have cared.
Four or five years later, however, pilot-wire relaying began to be sold. And there was in the City of Terre Haute, Indiana, they were going to put a loop around it with a lot of short lines, pilot-wire relays. The salesman from Indianapolis, Freddie Green, came into Switch Gear one day, and he was crying on my shoulder about Westinghouse. GE was going to offer the GMB, a relay they had developed for carrier current. They were going to use it for pilot wire. Instead of a carrier current signal, there'd be a pilot-wire signal. Green said, "Westinghouse has nothing to offer to compete with this." And I said, "Well, I have this system that I patented four or five years ago."
One weekend, a few months before, when I was testing some of the carrier current relays on our miniature transmission line in East Pittsburgh, I had taken the weekend and gone in and scrounged enough resistors and reactors to make up two-sequence networks and had tested it. It had tripped for all the internal faults, and it didn't trip for any external faults. Now I said, "It's completely undeveloped. If you want to take a chance on it, it's up to you." Well, he said, "What have we got to lose?" So we got Newark to put a price on it, and they priced it about half of what the General Electric one was. We still had a tremendous ratio of sales price to cost because the relay was only one little element that opened and closed. And all the rest of the brain was in the network. These resistors and reactors, which were very inexpensive, had nothing critical at all about them. The brains were all in the network, and it was just one little relay.
So it boiled down to a meeting with Joe Trainer who was the chief engineer of the Public Service Company of Indiana, which ran the Terre Haute operation, and Arbuckle, the relay engineer, who had practically promised it to GE. He had done this there was no competition. So now he had to go into his boss and explain. He said, "Here Westinghouse is offering this relay that only has one contact. If that fails we don't have anything." Joe Trainer looked him right in the eye and said, "All these years you've been wishing that instead of this whole mess of contacts you could have one contact that closed when it should and doesn't close when it shouldn't. And now that they're offering you exactly that, you don't want it. I don't understand." So the result was that GE didn't even get to use their GMB even for the pilot-wire application. It was like adding fire to the coals. Nobody ever mentioned it in the Relay Committee. I liked these guys at GE; they were all friends. But that's the fortunes of war.
That HCB relay is still widely in use. Now patents only last 17 years so obviously anybody could build it, and [certainly] in the last few years. But there's a tendency for companies not to just pick up another company's product and start building it. So it's still largely a Westinghouse product. And it also was worth many millions of dollars to the company in business. When I got back to East Pittsburgh, after my visit to Newark, the HCB relay came into operation.
Electrification Projects and Hoover Dam
Aspray:
Were you still in General Engineering at this time.
Harder:
No. During the Depression, General Engineering broke up, and let everybody go that wasn't a sponsor engineer for one of the districts. So about 1933, Pete West came from Switch Gear and said, "Ed, as long as I have a job, you do." I was handling all of the technical work for the Pennsylvania Railroad, in railroad electrification. We had made a relay study for them. The chief electrical engineer from the railroad, Harvey Griffith, would come over once a week and we'd discuss it with him and then we'd go on. For some months we studied all different phases of the relay system and came up with the system they use — even today — for the whole Pennsylvania Railroad electrification. Griffith and I were going to write a paper on it, but the railroad decided it wouldn't be fair to GE to have them cooperate just with Westinghouse on that. So they let me write the paper it's entitled "Pennsylvania Railroad." They use it as a bible; it tells all about their relay system.
General Atterbury was president of the Pennsylvania, and he was a very far-sighted man. He didn't believe that America was going to be in a depression forever, and he went right ahead with his plans during that period. Every two years he started a new electrification, all during the Depression: Philadelphia to Westchester, Philadelphia to Trenton, Trenton to New Brunswick, the New York Zone Philadelphia to Washington and finally out to Harrisburg. It was about 1934 by then. The Depression was, ending he went right ahead. We had a $10 million order for locomotives in East Pittsburgh, right in the middle of the Depression. It almost kept East Pittsburgh alive, Atterbury deserves a lot of credit for having the faith in the economy of the country to do that in the face of depression. Of course he got low costs; he had no labor problems all during the Depression. Everybody was glad to supply anything or to work for you. He got benefits, but he had to have the faith that this wasn't going to last forever. Well, when those electrifications were over, they studied coming onto Pittsburgh. The first time they studied it, there was $13 million savings. General Motors had come out with a beautiful diesel locomotive by that time. The last time we studied it, it was $1 million savings, and the next time there wasn't any savings at all. So that was where it stopped, at Harrisburg.
Then West put me on the Hoover Dam project. We had the switchboards for the first five units out there, and it was divided into three parts with a lead engineer and two others. I was one of the others. I had to design some of the switchboard for Hoover Dam, working on the enunciator systems and the load and frequency control and the terminal boards. I never saw it in operation until ten years later because they sent the lead engineer out there to live with it until it all worked. It was only on a vacation trip years later that I got to see it all in operation.
Aspray:
At this time had Westinghouse been able to keep on most of its work force, or were you unusual in being able to keep your job?
Harder:
Unusual. We all had furloughed days during the Depression, and many people were let go. Of course the shop was way down so this was engineering, shop, sales, every place. If it hadn't been for the Railroad, I certainly wouldn't be working for Westinghouse today.
I also worked on the voltage regulators for the Safe Harbor Water Power plant, and then went to Newark. It was after that that I was working in the relaying group in Switch Gear, and Monteith had become manager of Central Station Engineering. He got me back, and I got the Middle Atlantic District, which included Washington, DC. had Washington all during the war, so I had all the dealings with Captain Rickover there and the rearmament program before we got into the war. Then we lost 19 warships at Pearl Harbor, and I was in on the ships' service power supply systems for all the new ships there.
Aspray:
Would you talk in more detail about some of those activities?
Harder:
I'll first talk about the Navy activities. Before the United States got into World War II, they were beginning to get money for rearmament. The Navy yards were in a very bad state of disrepair; they had been engineered by people that knew absolutely nothing about electrical systems. I saw a list by the consultant that was working on it about how much the circuit breakers would interrupt and how much the duty was. The amount of the short circuit current was 10 to a 100 times what the circuit breakers were capable of interrupting. These had just been put in by civilian employees in the Navy. So they got money, and all of the Navy yards were re-stored to respectable standards of power engineering of that day. I was visiting Washington once a month, so I was in on a lot of the advice about that. Then after the Japanese struck Pearl Harbor and we lost 19 warships the Navy became very busy building new warships. They appointed Captain Rickover to head up Section 660, which was the section responsible for the power supply system. This was not the main drive but all the power aboard ship, called it the ship's service power supply system. Captain Rickover was in charge of that. One of his first actions was to come out to Westinghouse and try to hire some people that knew something about it because we had good engineers in distribution systems. Monteith let him have Phil Ross one of the better young engineers. A few weeks later Rickover was back for more people. Monty said, "Nothing doing. You're not making use of the man we let you have." Monty had been keeping track by telephone, and Phil Ross was down there just sitting at a desk not doing anything. Probably the lines were being chewed up before Rickover ever left Pittsburgh, I imagine. But I was visiting the Navy regularly, and Phil Ross was in on every meeting after that. He soon became the lead engineer, and he was so much more capable than any of the ones that they had then on the staff. He did a fine job. He later ran the Bettis plant here in Pittsburgh. He became one of Rickover's favorites.
Both Westinghouse and GE got all of their experts on that type of system. People on steam turbines, on governors, on generators, on voltage regulators, on switchgear. On every part of the ship service system we got our best people. GE did the same, and we all worked together. There was no anti-trust in those days; we all worked together. But the old ships didn't have the modern radars and fire control equipment that the new ships would have. So the new ships had a lot of sensitive electronic equipment that had to operate, even while this little power supply system was swinging gun turrets and raising elevators and doing heavy-duty jobs. The voltage must not drop so much that it puts the radars out of commission. So I was a general engineer, and so we got experts on all the components. I handled the systems part for Westinghouse. Sel Crary handled it for General Electric; he was in a bad automobile accident later and hasn't been heard of for a long, long time. He was a darned good engineer. Monteith kept in close touch with Rickover all during this time, guiding the whole program. There were many all-night tests, both in GE and at Westinghouse, to try out phases of the system as it was being developed. In the end we had the battleship "New Jersey" in the Philadelphia Navy Yard for three days, when it was urgently needed out in the Pacific. We had it for three days to test all these systems, to make sure that things that we had put into them really worked. So later when I heard about the "New Jersey" being out in the Mediterranean off Lebanon, I said, "Oh, there's an old friend." I've been aboard that one. So anyway that all worked out.
The relationship between Monteith and Rickover that started there was very important later when Westinghouse got into the nuclear field, because Monty knew Rickover well, and he knew that what Rickover set out to do, he could do. They had a lot of confidence in each other. So later when Westinghouse undertook to build the reactor for the "Nautilus," first nuclear submarine, we had a banker for a president. Price was the president of Westinghouse, and the senior vice president was a businessman. A previous president was a legal man. Monty was the only engineer giving advice. He recommended going ahead and building this power plant for Rickover; which we did. We later built the Shippingport plant, the first land nuclear plant, and after that a whole flurry of nuclear plants. It's important that this early contact between Monteith and Rickover is where the mutual trust and knowledge of each other came about. So years later when it came to building this nuclear stuff it all sort of fitted into place.
The most important aspect for was working with the Navy. And as I say, I handled the systems end of it, putting it all together. All of the characteristics that these fellows could get in their turbines and generators and governors and voltage regulators, I put it all together as to what overall performance you could get. In the process I developed a regulating system simulator, an electronic device. I had an assistant that built it, but I designed it. After the war we incorporated that into the pilot model of the Anacom. I wrote my dissertation on the general solution of the voltage regulator problem by electric analog computer. It was all based on the work that I'd done with the Navy during the war. Of course that's where it all started.
Aspray:
That simulator was used effectively in your work in Washington?
Harder:
Yes. In other words, with the characteristics you could get in all these machines somebody had to calculate what the voltage would drop to when you got sudden loads.
And I used that simulator for part of it, I not all of it. I did some hand calculations, and when wanted to make a general study I'd use a simulator. You could cover a whole lot more ground in the same length of time than you could calculating it longhand. Of course, when I wrote my dissertation I covered the topic still more broadly far beyond just what the Navy needed. It was sufficiently mathematical, however, that Pitt decided that my degree should be in mathematics.
Graduate Work & Voltage Regulator Theory
Aspray:
When did you begin your graduate work?
Harder:
In 1945, two professors from Pitt, Professor Dyche and the head of the math department came to and said that they would like some of fellows that have all the credits needed to go ahead and get your degrees. I said, "Oh, I'm too old. I'm twenty years out of college. I'm teaching and doing other things. I'm too old to get a degree now." And they said, "We don't think so." And, you know, I thought about it and I said, "Well, if you don't think so, I don't know why I should." So that was when I decided that this would be my thesis project: general solution of the voltage regulator problem. I had the professors from Pitt out to see the apparatus in East Pittsburgh. Because they were already good friends of mine from way back, there was no question about any examination. It was more or less cut and dried. They were on my side right from the start. And so, I took a year studying group theory with Professor Culver, they wanted me to know something about the broad field of math if they were giving me a math degree. I have told people that every bit of math that I studied at Cornell or subsequently, I have used. I didn't just take it; I took it because I needed it. "Well, what did you use group theory for? "And I say, For exactly what they told me: to get a broadening in mathematics.
Aspray:
I wrote my master's thesis in group theory.
Harder:
Well, then you can understand better than most people I talk to what I'm talking about. Certainly there was no specific engineering problem that I can point to group theory with, but I do know what mathematicians work on, and I can see all the problems there are.
Aspray:
How did you handle all the course requirements for the Ph.D.?
Harder:
I had taken them. In other words, up to the master's level the engineering school gave 15 credits, and in 6 more I had a master's degree. Then there would be a math course in differential equations or in geometry or something that I wanted to take. Then high frequency techniques appeared during World War II. Nobody knew anything about that before. They used a book on ultra high-frequency techniques by Brainerd, who was at MIT then. I went back to school and studied that, and I used it. It's involved in carrier current and radar and lots of the things I got into.
That takes me through World War II activities and extends to my getting my doctor's degree. Really the thesis was based on the work that I had done with the Navy during World War II, studying regulating systems. I realized from those studies that it would be possible to make a general solution. I saw that there weren't many main variables but that you could vary them and see how a regulating system worked. That became very valuable later because I realized that a two-delay system is always stable. A three-delay system can be stabilized with damping transformers and devices, but it can be unstable, too. Westinghouse was in trouble with steel mill regulators, paper mill drives, generator voltage regulators, etc. I can see now it was all due to three-delay or four-delay systems and probably could have been stabilized for one condition; they would just try to change to another condition. They become unstable and create all sorts of trouble. They cost millions and millions of dollars, and the poor guys in the design departments were rebuilding their exciters and things that had nothing to do with it. They thought their characteristics weren't good enough; the whole problem, however, as my studies then showed, was that they were multi-delay systems.
When magnetic amplifiers came in, you could make them as fast as you wanted. If you excite them with 400 cycles, only a few cycles delay, but they're very short cycles. So prior to that we had rotating machines. All during the war GE had the amplidyne and we had the two-stage rotator. Those were the things that we were trying to make work in regulating systems. With magnetic amplifiers, however, we were the equal of anybody because we could essentially make the first delay zero. Then it was only two delays, and it was stable. It could be made much better but it would never get unstable if there were only two delays. If we had only known that ten or fifteen years before, we could have saved millions of dollars. But it required a general study to give a general concept like that. They didn't look at things that way; they were just building these regulating systems the best they could. Research solved all equations for the steel mill at Gary, but only for one condition. They wanted to roll thicker steel or thinner steel or speed it up or slow it down, and it would be unstable. Then it would be back to the poor engineers in East Pittsburgh. If they'd only known that the only solution was to have a two-delay system.
Aspray:
Did this result of yours have practical significance for the company?
Harder:
Yes! When magnetic amplifiers came in, we solved all those problems. We used these magnetic amplifiers on steel mills, paper mills, and generator voltage regulators. Up until then, if I must admit it, GE's amplidyne was better than are rotatrol, but it still wasn't perfect. The magnetic amplifier, however, was better than either one because you could make it fast just by building a 400-cycle set to supply it. Essentially you could eliminate one delay. We got all those problems solved because we used a fast magnetic amplifier to replace the first delay in all those systems, making them all stable. It's amazing that if we'd only known that years before, and had a magnetic amplifier years before, millions of dollars could have been saved.
Aspray:
Would the problem still have been solvable if you had not had a magnetic amplifier?
Harder:
Well, let's say you wanted to roll different thicknesses of steel at different speeds. If you were willing to limit it to four, I suppose you could make four sets of constants that were stable and switch the feedbacks and so on as it changed. Research really led them astray in solving for just one condition and assuming it would be all right for others. Clint Hanna was the head of that group in Research, his experience had been with systems that didn't change. He got a presidential award for the tank gun-stabilizer that was used successfully at El Alamein. That, however, was a mechanical system, and it was fixed. The stabilizer kept that gun pointing right at the same place even though the tank wobbled all over the place. So our troops were able to shoot the other tanks while they were moving.
That was the sort of experience he had when he was giving advice to a manufacturing division. But their problem was different. They had a variable system. It was always changing, whereas the tank was one thing. You had one gun, it had certain characteristics, and this thing had to keep it fixed in position. I don't know if Clint ever knew this either. These things become clearer as you get older and you look back it and see why all this occurred. The Anacom was finished in '48. While the steel mills were having all the trouble at Gary, we had the system set up on the Anacom and also set up actually in Gary; we were working together on it. That was before I had made this general study. And I don't think I was aware then that if I had — Well, for one thing, we didn't have the magnetic amplifier yet. So we were just trying things to get out of trouble. Later the general study showed that you're out of the trouble if you could get it to be a two-delay system.
Aspray:
When did you have that result?
Harder:
I had it before the Anacom was finished in '48. We didn't have a magnetic amplifier anyway. After world war II, knowing that our rotatrol was not as good as the amplidyne, I was afraid the rotatrol might not make it on the ships. I concocted a system using saturable core reactors, that I called a balanced, biased saturable core reactor amplifier. If we didn't make it with the rotatrol, we had this as a back-up. We never had to use it, but I wrote a paper on it after the war. Gordon Brown read the paper, and we put all the constants in the paper. He had his students build the thing; then he offered me a job as professor at MIT, which I turned down.
Aspray:
Tell me more about your career in the late 'forties.
Harder:
A lot of the work I did during that period had to do with the Navy. From 1938 until 1946, I was assigned to the Middle Atlantic District including Washington, DC as a sponsor engineer. In the latter part of that period, we were visiting the Potomac Edison Company, which is headquartered in Hagerstown, Maryland. It's a system about 100 miles wide, maybe from Cumberland to Fredericksburg, about 50 miles north-south from Pennsylvania Turnpike down into West Virginia. Their headquarters was in Hagerstown, and the salesmen from Philadelphia would take me down there occasionally. They had one good technical man, and they had an older manager who had previously run the streetcar from Pittsburgh to Butler. Apparently the company had put him down there to run this power system, and he did a commendable job. But they were having a very unusual problem; at nights and light-load periods the whole outskirts of their system would have a very high fifth harmonic voltage content. It would get 10 or 15 percent fifth harmonic voltage. On a capacitor, 100% of 60-cycle voltage will produce 100% current, but it only takes 20% of fifth harmonic to produce 100% current because the impedance of the capacitor's only one fifth as much. The older capacitors didn't have enough reserve in them to handle that much extra current, and consequently, the switchboards were burning up. They took me up to the Pennsylvania Turnpike and showed me some of these stations where the switches were all blistered. They were very confused. As I told you before, I'd had a lot of experience with harmonics. I knew harmonics better probably than anybody because I'd had this dual experience of telephone interference and building filters and teaching harmonics, and the Pennsylvania Railroads problem with harmonics.
They didn't have any rectifiers, and the only source of harmonics is the transformers. If a transformer has to produce a sine-wave voltage, it has to have a sine-wave of flux. And the current is not going to be sine-wave. It excites it. It's going to have a lot of harmonics in it. I knew a couple of different equivalent circuits of transformers and correctly represented them. I knew that up at Sharon we had used several kinds of iron in transformers. Just from the characteristic of the iron, I could calculate what these harmonics would be. It wasn't too different for any of these five different kinds of iron they used. I assumed that GE's iron must be about the same and that this equivalent circuit would probably work for all transformers. So I set up the whole Potomac Edison system on the AC calculating board at 300 cycles. I started with the equivalent circuits of the transformers, which is where the harmonics were coming from. I represented all the loads; mostly guessing at them. The starting current of a motor might be five times normal. Well, that's when the slip is 100%, when the motor's standing still. The fifth harmonic is the same. So motor load is roughly 20% impedance at the fifth harmonic. I guessed at how much was lighting and static loads, and how much was motors, I put the loads at all of these stations. Then I ran it on the calculating board, and they were right at the peak of a resonance. If you added capacitors to the system you passed the resonance point.
So I told them that if they would install a 2,000 megawatt capacitor at Winchester, Virginia, it would solve all their problems. Nobody had ever done this before or has done it since, so far as I know. A 2,000 KVA capacitor was at least ten times bigger than anything they had ever installed before in the way of a capacitor. They'd installed 100 or maybe 200, but not 2,000. So they asked GE about it, and GE wrote them a letter stating that it wouldn't work, and recommending against it. But the technical engineer there, was a radio ham, too, and knew a lot about electricity. He knew enough that he could go to a station and filter out the fifth harmonic and put it in an oscillograph and measure it. He could get the right circuits set up so that with an ordinary oscillograph he could take a picture of what the fifth harmonic was. He saw my calculations and asked why I concluded this. He believed my reasoning, and they installed the 2,000 KVA capacitor at Winchester.
There were tests, and it did exactly what I said it would do. And then I told them if you put a coil — a reactor — in series with this capacitor and tune it to the fifth harmonic, the current that it will draw will be within its capacity. I had measured on the calculating board how much this current would be, and there wasn't too much. They could actually tune it, so they called it a filter. They put this capacitor on, and they got the benefit of it for power factor correction. It actually short-circuited the fifth harmonic at Winchester. And at Frederick, 50 miles away, it brought it down to half. There was a fellow from the telephone company there to witness the test, and he wrote me after he retired. He was so nice at that test; nobody had ever tried it before. I wouldn't have tried it if I hadn't had all that experience with harmonics. And I had confidence in my calculations. So then Wilbur Feaster, the customer's engineer, told me several months later: "Well, GE came around and they asked for their letter back. And he was a nice guy, so I gave him his letter back."
Anacom
Aspray:
Tell me a bit about Anacom.
Harder:
Well, in 1946 I was made consulting transmission engineer and put in charge of lightning field research.
Aspray:
Does that mean you had a certain group of people working for you at this time?
Harder:
I didn't have any group, other than consulting. But I had a lot of freedom to travel all over the country. I made several speaking trips of five and six weeks at a time that I couldn't have done if I'd had continuing responsibilities. In 1945 or early '46 there was a fellow by the name of Gordon McCann with Westinghouse, and he knew about several of the earlier analog computers that Westinghouse had built. Of course he knew about the DC board. Then Evans and Monteith had devised an electrical transients analyzer with a rotating synchronous switch. It's driven by a synchronous motor, and has handles with which adjust the closing time each cycle. It goes around ten times a second. So you can set up a transient, repeat it ten times a second and display the result on a cathode-ray oscillograph. That was the way the Anacom worked. Later on at Anacom II, that was replaced by an electronic synchronous switch. It wasn't a big mechanical device anymore, just the electronic circuits that drove mercury-wetted relays to open and close the circuits. That rotating switch took a lot of maintenance.
Well, McCann knew about this electrical transients analyzer that Monteith and Evans had worked out to study electrical transients on power systems. Then McCann and Harry Kriner had developed a mechanical transients analyzer, which used essentially the same equipment but used electrical analogs of mechanical systems so they could study the transient torques in turbine generators and rotating machinery. Kriner was a mechanical engineer. McCann, was an electrical engineer. They combined on that. I had developed a servo analyzer, working with the Navy. I had it in East Pittsburgh, and I had used it. McCann proposed that we combine all these things and make a general-purpose analog computer with a lot more elements, and flush it out so it would be really a big full-scale device. Then in 1946 McCann left and went to Cal Tech, and I became consulting transmission engineer, and I took over this work. We had already built a pilot model that year. We just threw it together using pieces of equipment. That's the one I did my doctoral dissertation on.
So based on the good success with that pilot model, we got $500,000 — which would be $5 million today — to go ahead with the full-scale model. McCann had left, putting the entire operation in my hands. It was agreed to build a second set of parts for Cal Tech at very low cost so McCann could set one up out at Cal Tech also. I had a number of people working on this at that time, developing low-loss reactors. You could get molypermalloy iron for the transformers. Reactors had air gaps in the iron, but the iron loss had to be very low in the reactors, because we were going to use it over a range of frequencies from a few cycles up to 1000 hertz. Well, we could have just collected a lot of equipment, then add leads that connected it together. But I made the decision that we were going to build it like we built the AC calculating board, so it looked like a real machine. If we hadn't done that, we wouldn't have it 42 years later. That was an important decision to make. Then I had to make decisions about all the components and the wiring and the instrumentation and everything about it. But by 1948 it was in. The costs ran way over what had been estimated, and my boss was furious. I said: "Well, I didn't estimate it. You put me in at the midpoint, and I didn't have any control over it. These things were all ordered when I came. You can fire me if you want to, but there's nothing I can do about it."
Aspray:
What was the cost?
Harder:
Five hundred thousand dollars.
Aspray:
And what had it been estimated to be?
Harder:
Oh, maybe $300,000. You see he had to go downtown and explain overrun on his budget. I was never in a position where I controlled those costs at all. Maybe I should have been in a position if I'd had more experience. Instead we went right ahead and did a first-class job. Got these first class-cabinets and made it all in first-class style.
Aspray:
Were there people in other companies building this kind of equipment?
Harder:
GE never built one. They undoubtedly in the laboratory put assemblies of reactors and capacitors and so on together to solve specific problems, but they never built a computer. So most of the experience was our own. We had a small scale electrical transience analyzer and we had the mechanical transients analyzer, so the analog idea was there. Now even the electrical transients analyzer is just an analog because you don't have transmission lines. You have reactors and resistors to represent them. It's an electrical analog of an electrical system, and there's not much difference if you have an electrical analog of a mechanical system. The only requirement is that the equations be the same in the actual system and in the analog. That's what makes it an analog. So we had the know-how in these two previous machines and in the servo analyzer, which I had built. That's all we needed. It became the most important computational tool in the company immediately, and it was used by all the divisions.
Aspray:
What kind of training did you need to be able to use it effectively?
Harder:
We had to have some way of checking everything he did. They had to know that their answer was not off by much; we had to have some other way of looking at a problem. One of the first big problems that we worked on for South Philadelphia was the Ullahoma Wind Tunnel. It had long blades, and for the design of these blades you had to know at least the first four modes of vibration. Nobody had ever been able to calculate anything above the second. But on the new ANACOM we could set up the analog of the blade.
We were supposed to share the development with CAL Tech. Well, the only thing we ever shared was actually quite valuable. They had a fellow by the name of McNeal, who is now one of the principals of McNeill-Schwendler out in California. He had set up the analog of a cantilever beam. He had represented it in about 10 sections, made the analog for each one, connected them together and then checked it. Now a regular cantilever beam has a closed-form solution: you can do it mathematically. So for that case he could check how good the analog was. He could set up the analog, and he published this. There was good correlation between what you got and what the closed-form solution showed it actually was. So we took that and extended it from a regular cantilevered beam to a tapered beam and a twisted beam, which is a blade. We hoped that they would be equally good, but there was no closed-form solution there to rely on. So we just assumed that since the other was good, that this was good, too. They used it for the Tullahoma Wind Tunnel blades. Once we got the thing set up, you could apply an audio oscillator to it and put different frequencies on in minutes and measure what the modes of vibration and the natural frequencies were very easily. Then after you built the blades, you just hit it with a hammer, and you could tell what the natural frequencies were after it was built and in place. So after it was built, we got it checked, but not during the building process. And it was good enough. That was in 1948. Just the next year we had a digital computer that we calculated one or two cases through, verified that this was correct to within 1 or 2 percent.
Aspray:
What other kinds of calculating equipment were you using at Westinghouse at the time?
Harder:
When I say in 1949 we had a digital computer, that's what it was. It was a 602A. A machine that had a board that you could wire to do half a dozen things. It would read a card and make little calculations on the numbers on that card and punch the answer back on that card. We'd set up this blade calculation in 10 steps. By continually resorting the cards and reentering them manually, and trying different frequencies, you'd find out which ones came out zero as those were the natural frequencies. And as I say, we had that machine available in our accounting department in East Pittsburgh in 1949 — only a year after we built the Anacom. So it wasn't too long before we had some digital check of this calculation. By 1950 we had a Card Program Calculator (CPC). There you could have a program as long as you wanted, a whole deck of cards. And the board could do ten operations: floating-point arithmetic, trigonometric functions, roots and powers. The fellows out at Northrup Aircraft showed IBM how to do this with machines they already had. The CPC had a purely mechanical memory of 36 numbers, 10-digit, numbers. It could pick out any of the 36 numbers and do any of these ten operations and or put the results back in memory. And so that greatly simplified it. But we did it first with this simple, punch card program and card-reading machine. In another year, however, the little drum computers came out, and then you could store the program, too. So it wasn't long after the start of the Anacom that digital computers arrived.
Aspray:
Yet there was still a place for the Anacom.
Harder:
Yes! With the Anacom, once you set up the analog, you could measure everything immediately. You didn't have any waiting time. There were a lot of problems that didn't lend themselves to this interactive procedure that the blade calculations did, so I had what I called the Anacom Manual, in which we had a key sheet for each problem we solved. We asked questions such as: What would be used for the analog? What were the inputs and outputs? How did it work? That was our manual for the Anacom. The problems that we could solve included: oil flow problems, lubrication problems in bearings, and those that were represented by either differential equations or diffusion problems.
Aspray:
Could you discuss the situation with the analog versus the digital computer?
Harder:
Well, the Anacom was built as a general-purpose analog computer, and it had problems: electrical, mechanical, fluid dynamic, diffusion. One by one they got displaced. By the time the CPC emerged, we were just starting to do network solutions and invert big matrices. That was just the beginning. By the time of the invention of the 704, which was the first big tube computer that we had, you could do pretty respectable network solutions. But even in '59, when I went to in Europe to do that we were still using both. For example, for a stability study you might do the network solution on an analog computer and the stability study on a digital computer. The result of the network solution is adaptable to the analog computer. Then you would use that solution in the stability study of the system. So during this period from 1950 to even in 1959 we were still arguing about which was the best way to do it.
In World War II the Mark-IX gun computer was used in Europe at Anzio for shooting down planes. Then the proximity fuse was developed. The Germans shot 3,000 V-1 missiles at Antwerp, our main base in Belgium, and 1 percent got through. With the proximity fuse you had to come within 50 feet, but they shot them down with an average of 18 shots per missile down. Well, after the war there was a whole rash of commercial versions of these. These Mark-IX gun computers were essentially the electronic differential analyzers. Little differential analyzers. Everybody had one. Lockheed had one, Reeves had one. There was a Pace computer. We got a Pace computer for the Analytical Department, and they were much better equipped to handle these regulating system problems than the equipment we had on the Anacom.
Beyond 1959 into the 'sixties and 'seventies, A man named Rideout developed a program called Digital Analog Simulation (DAS), and a program called Modified Integration Digital Analog Simulation (MIDAS). With that essentially you could put a Pace computer on the digital computer anytime you wanted. From that time on I would say the electronic differential analyzers went out of business. Because you could do them on any digital computers with Rideout's programs. Until today it's only these transient and lightning and oscillatory problems and nonlinear problems of electric power systems for which the Anacom is still used and probably will be used for a long time.
Aspray:
Why are those problems not suited to a digital machine?
Harder:
They're transient. I'll give you an example. When the 701 was first available in New York, I wanted some problem to try out on it. A traveling wave going along the transmission line travels with the speed of light. With the part of it that's above the corona level of the line, the capacity is about four times greater, and the inductance is no different. So it slows that part down to half the speed of light, and this wave comes apart. So I had solved this problem on probably the 650 in East Pittsburgh. I thought then with IBM's offering the 701, it could be so much faster, I could take a lot finer steps. I could find out how good my solution was because I could make the steps all so much finer. My solution wasn't too good however, because it blew up. In other words, you cannot represent a ramp function on a digital computer. You have to have steps, and those steps have a frequency. They may excite anything in the system. It is difficult because the systems are so vast; they represent two or three hundred lines on the Anacom or these big analog machines. With traveling waves, it's hard to represent one or two sections, let alone a hundred.
Gregory Vassil was senior vice president for planning of American Electric Power. He used our Anacom for a problem where if you only had lines coming in, you could get in trouble with oscillations when trying to start up the plant. The capacity of the lines interacted with the frequencies of the things you're trying to start, and this could blow up. When the office of American Electric Power was in New York, he used to come out to East Pittsburgh every few months to run studies on our AC calculating board. Through that we got pretty well acquainted. He keeps me informed of things he does. The trouble is it's just entirely too complicated to try to represent this: the analog is such a complex system. Perhaps you could do it by building blocks and get a building block for each part, but I would be afraid that you would have to use steps in the process to do it digitally. There is no ramp to go from one to another. On the analog, however, there is. All kinds of weird things happen in power systems, and they can study them all on the Anacom: they just have to represent it and study it. There are now several machines like the ANACOM in the United States, and there must be several more throughout the world. More power to them if they can do those things digitally, but I tried my best and I can't get to first base.
Aspray:
Was the Anacom made available to outsiders?
Harder:
Yes. There's one in Pennsylvania someplace that's owned by a consulting firm. At first, when I had the Anacom, it was a profit center, charged for everything. There was no free time for Westinghouse on it.
Aspray:
Did you have to set your rates higher so that you didn't have too much business?
Harder:
No, and yet it was quite well used. You'd think it would either be under-used or that there would be a waiting line. I don't recall such.
Aspray:
Was there a service provided to help people use it?
Harder:
Oh yes. We either had the know-how or we got help from Research. Once we got a class of problems running, less skilled people can put in the inputs and make use of the outputs. But up until that point, why, we provided the know-how.
When we first started the Anacom, some of the fellows had a few lessons to learn and made mistakes. But by the time that we had the Boemark Missile Project and Sunnyvale had the contract to build the launcher, I had fellows that I could trust. Maybe you're trying to calculate that the clearance between a missile and the launch tube. We handled that all in East Pittsburgh for Sunnyvale with fellows that I could trust. The key to it was they wouldn't make any big mistakes. By that time they had become skilled in this business of "you're pretty damned sure it's about so!" Maybe the last little bit you depend on the computer, but they didn't make any big mistakes. We had that record of never getting any department in trouble for the 20 years.
There were, however, some fellows I just had to get out of the department because they didn't have the kind of a mind that could sense whether it was their imagination or the truth they were depending on. They couldn't separate truth from imagination, and you can't do that with an analog computer. It has to all be correct and factual.
Aspray:
Who was allowed access to the machine?
Harder:
Utilities all over the country would use it for the lightning and switching transient studies on their systems. Now we had an Advanced Systems Technology Department that was a profit center; they didn't give away their services. Westinghouse took on the contract to develop the first 500 KV system for Virginia Electric Power Company VEPCO, and then for Allegheny Systems, here around Pittsburgh. These were the first systems where it was determined that switching transients were an important parameter in the design of the systems. And where circuit breakers were actually developed with resistors to cut in on closing to cut down the switching transients for economic reasons. So in that case Westinghouse was directing the projects. We were getting paid for it, acting as consultants. All over the country, however, there would be people that would have problems of this kind that would have to go to one facility or the other if they were going to do real design work on the switching transients and lightning performance of their systems.
Aspray:
Would you give access to a company like GE?
Harder:
No, but they wouldn't come to us, of course. Nevertheless, GE has one now. The people that run this Anacom wrote to tell me of other systems that existed. They were pretty sure CESI has one in Italy. And I know there's one in England. I know there are others, but we have no complete listing of them.
Aspray:
Was it a great technical challenge to build the Anacom?
Harder:
Well, I think that after we had the pilot model and it worked all right, and we were able to solve all the types of problems, it was simply expanding into a 100 elements instead of 10. The main challenge was how to make such a big machine without running the capacity of the wiring up so high that it vitiates the solutions. We had a number of very good engineers building these reactors and transformers with low loss, using molypermalloy iron and 20 pounds of molypermalloy in a transformer. This was unheard of, but we had to get the loss way down to do this kind of work. In that sense it was a challenge. But we looked at it: Well, we're just going to do the best we can with the fellows that we have. With an analytical department in a company like Westinghouse, the very best engineers seem to gravitate to it. And we certainly had our share of the very best. Some of them were working on the Anacom at that time, designing these reactors and transformers.
Aspray:
Was the Anacom described in the public literature?
Harder:
Yes. It was described in a paper about the AC calculating board, which is certainly one of the predecessors, except this was run at 400 cycles. We were solving fixed-frequency problems. This described that one predecessor, which was the electrical transience analyzer. As time went on there were a lot of new techniques developed: things that you could do once you had this facility. So Carleton and I wrote a paper describing a lot of these new techniques. Then I gave a paper in Brussels describing all the Analytical Department; there were six computers at that time — two analog and four digital. This demonstrates a typical diffusion problem. Heat flows out through a rotor, which you can do on a DC calculating board. A resistor network must be set up. They used that for nuclear reactor design at first — diffusion of the slow flux and the fast flux in the reactor. Later digital calculations were used to interconnect the two. How much fast flux would come from the slow flux at each point in the network. It goes through a diffusion and then a digital calculation, and then another diffusion. Finally you get a steady-state solution. Later they had complete codes for the digital computers for lifetime burnout of a reactor or any part of that you wanted to study.
After it they got digital input and output, meant that you could program what you wanted to run on the Anacom, and it would go ahead and go through all these cases without your sitting there. That, however, was after my time.
AIEE Computer Committee & Merger
Aspray:
Let's continue the history of your career.
Harder:
Well, largely because of my analog work, I had been involved with the AIEE Computer Committee. The AIEE was organized a general membership with committees dealing with different things of a limited number of people. The IRE, with which they merged later, had professional groups in which large numbers of people were involved in each activity. But I came up through the AIEE, up to the point where the merger took place, and I was chief officer of the merger between the AIEE Computer Committee and the IRE professional group on computers. And some of them said: "How are you going to merge a committee with a group? One is like 20 people and the other is thousands of people." Well, we did it. I think the answer was that there was goodwill on both sides. Everybody was anxious to get the best features of both societies built into the resulting society, and they realized it wasn't going to be perfect. They therefore voted to merge the AIEE Computer Committee with the IRE group. There are still a lot of committees in the IEEE. I visited the Relay Committee, for example, in Williamsburg, Virginia, this year. It's composed of hundreds and hundreds of people. The Relay Committee is a huge organization with all sorts of subcommittees.
Aspray:
In the AIEE committee, how were people chosen to be participants?
Harder:
They were chosen by people that were already on the committees or the organization knowing that they were the knowledgeable people and that they should be there.
Aspray:
What was the function of the committee?
Harder:
They approved sessions and papers and publications. There were papers on procedures, but mainly on technical matters. There was a lot of committee work having to do with the standards, too. I suppose that in the IRE the professional groups somehow managed to do similar things.
Aspray:
How long had the Computer Committee been in existence?
Harder:
I think it probably all came after World War I. [The 75th anniversary was celebrated in San Diego in January 1992.]
Aspray:
And was there representation for analog in AIEE?
Harder:
Yes. The Computer Committee had an analog subcommittee, of which I was chairman at one time. Frank McGinnis from General Electric was also involved.
A common thread to all this is that there was always an Electronics Committee, a main committee, and it started subcommittee after subcommittee. It started a Computer Subcommittee of the Electronics Committee. It started a Magnetic Amplifier Subcommittee. All of those became main committees later, but their germination point would be either the Research Committee or the Electronics Committee. They would know the people that they would get to head these other activities. So the Electronics Committee started many other committees over the years.
Aspray:
It seems to me that this AIEE Computer Committee is a microcosm of the engineering world. It would be interesting to look at how the competition between analog and digital played out in that committee as maybe a representation of what was going on more widely.
Harder:
Well, I'm sure that my initial representation on the committee was due to my analog work. Yet, some at that time were certainly digital. For example, an IBM representative wouldn't be involved in analog at all.
The American Federation of Information Processing Societies AFIPS was already formed. The Association of Computing Machinery was already formed. The IRE professional group on computing was operating, and the AIEE Computer Committee was functioning. They joined in the JCC — Joint Computer Committee, and ran the East and West Coast Joint Computer Conferences for a number of years until they federated and became AFIPS in 1960. In '62, was the first International Federation of Information Processing IFIPS, Congress in Munich. I remember giving a talk in Toronto about that time describing the world-wide cooperation in these fields and the cooperation that was going on in combining the AIEE and the IRE into a single society — with really no ground rules except the good faith of everybody that was concerned. Pat Haggerty, from Texas MCI, was the president IRE. He was really a strong man in engendering all this good faith and goodwill. He just understood that these groups — as disparate as they were — they were going to get together and form a great new organization. This was all brewing in the time before '63 when it finally culminated.
Aspray:
Do you recall if there were any major differences in outlook about the two organizations towards computing?
Harder:
I don't think so. I think some of the ones that were most heavily involved were in all of them. The fringe ones would be involved in one field or another. The point I made is that while this looks like merging a committee with a professional group, really the committee is representing all of the AIEE interests in that field, which is far beyond that committee. It represents all of the members of AIEE that are involved in computing; and that's where it centers, that's true. It's organized differently, but actually the number of people you're combining is not all that different in the two organizations.
Presidency of AFIPS
Aspray:
How did you get involved in professional activities in IEEE, AIEE, AFIPS and IFIP?
Harder:
Well, I had been chairman of the AIEE Computer Committee and actually handled the merger meeting with the IRE professional group on computers. At that time I had graduated from chairman of the Computer Committee to chairman of Science and Electronics Division, which included electronics and computers and a lot of high technology in the AIEE.
In 1960 IFIP had been formed and Mike Auerbach was president; he didn't want to handle all the U.S. activities at the forthcoming congress in Munich. He wanted to get somebody else to do that. So they proposed me. I had been chairman of the AIEE Computer Group, and now I was influential in IEEE. And so they asked me to handle the United States affairs at Munich. And I readily agreed with this because it was an opportunity to get acquainted with all of those leading computer engineers in the world. My boss went right along with it even though it meant going to Europe four times in the next two years. He agreed that this was well worth it. So then that all worked out fine.
Then it came up as to who was to be the next president of AFIPS. Well, there were three societies involved — ACM, IRE and AIEE. And they rotated the presidency. And it was AIEE's turn. So somebody really had to be selected from AIEE to be the next president of AFIPS. Apparently they decided that because the next congress was going to be in New York and all these Europeans that I had got acquainted with in Munich were going to be over and we were going to be hosts, that I was the logical one to be president. And so they sent a representative around and asked me to be the next president. And that looked entirely different to me. And I said, "Oh, no. One of you fellows that are in the computer field ought to be the president. It would mean so much to you. And I'm really in the electrical engineering field. I'm not in the computer field. We use computers — we use them a lot; we're big users of computers — one of you fellows ought to be the president." So I just left it at that. And about two months later, they came back again and said, "We still would like you to be president." Well, with that kind of support and their obviously seriously wanting me to be president, I couldn't turn it down. So I became president, and I guess I was reasonably successful. That's my distinguished service award from AFIPS hanging up there mainly because of being president.
Aspray:
What were your responsibilities as president?
Harder:
Well, of course we had board meetings. We had to decide where we were going to hold the East and West Coast Conferences. And we had a lot of legwork done. Charlie Asmith a staff member did a lot of the legwork on that. We had to make the decisions, on the various committees, and decide what to do with the money we had. I remember one tight decision. At International Congress an Information Processing ICIP Organized by UNESCO, part before IFIP was formed Congress '59, they had a large supplement the New York Times printed about computing. They got a lot of important people to write for it, and we got a lot of ads that paid for the whole thing.
When I was handling the Congress in Munich, I needed somebody to handle exhibits. Exhibits were going to be a big problem, and he had to work with the international people on the exhibits. And it was a big job. And I consulted Werner Buchholtz, and he said, "Well, Don Thompson, he's a mathematician, but he can do anything. He's an amazing guy." So I called Don, and he took on the job of exhibits for Munich. He was way out of his field, but he was doing a good job. He was very idealistic; he had an idea of what these exhibits should represent and all about it. And it was going to be his way. When it didn't go that way, why, he called up. He wanted to resign. And I told him, "Don, you have to learn to work with these international people. It's not all going to go your way. You have to compromise and work with them."
Well, before it was over he was that close to Billings who was the German representative over there, doing a lot of it. And he just did a wonderful job. So when I was president then, I needed a chairman for the Education Committee. And again I went to Don. Don characteristically took on a pretty big job. He decided to have another New York Times supplement. Well, times weren't as good then, and it was much more difficult to get ads. In the end, that was the best they could do — AFIPS was going to have to put $20,000 into it. I figured this money comes easy; We get it from exhibits. If we give some of it back to the exhibitors in the form of a New York Times supplement to help out their cause, then that's legitimate. But not all of my board agreed with that. They split six and six. I voted for it because Don had done so much for me. He was doing his best to do something here, and I wasn't going to let him down if I could possibly help it. So we had another New York Times supplement, and it was all right. I don't know if it was really necessary, but I'm sure the way those companies were putting money into exhibits then that we were getting several hundred thousand dollars from these exhibits. To give them back $20,000 was a reasonable thing to do. But, you know, a lot of the members of the board looked at it just as if it was money out of their own pocketbook. They were spending it, and they considered the $20,000 was too big a price to pay for the supplement.
Aspray:
What about other professional issues that came up during your term as president. It was a time when people were starting to develop computer curriculum. There were manpower issues. Did any of those have any real force during your period as president?
Harder:
Well, there were problems. In the computer field there were a lot of crackpots. And the job of a president of a society is to make use of those fellows and get all the good ideas they have, work it out so they get used, and get around all the crazy things that they try to do that you just can't do. And so Norbert Wiener died, and they formed a Norbert Wiener Society. And they applied for membership in AFIPS. Well, this was no more a technical society than the man in the moon. This was a worship society: they worshipped Norbert Wiener. Claude Kagen, who proposed the society, one of the ones that I had my hands full controlling. He was always wanting to do something crazy. This one was going to be awfully embarrassing because the board was not about to vote the Norbert Wiener Society into AFIPS. It was going to be a blanket turn-down and be very, very embarrassing. So I finally talked Claude into withdrawing the nomination.
A lot of other things came up. There were different ones that had crazy ideas. My job as the president was to use all the good ideas and get around the difficulties one way or another. If you can't talk them out of it, buy them off or do something with them that makes it work out. Well, they were all my friends in the end. I didn't make any enemies doing all this. But it took some doing, I can tell you. Now there were a number of other societies; some did get in. The Linguistics Society got in. The Society for Computer Simulation did, also. But it got to be quite a string of societies. I don't know what they all were because a lot of them came in after my time, and I didn't keep up with it all. At first there was a problem that the surplus from these exhibits was divided equally among the three sponsors. Well, when you start taking in other societies, do they get in on the goodies? You know. That was finally resolved, I think, during Paul Armour's term. I made the recommendation when I finished my term: "For God's sake, Paul! Separate everything AFIPS from the Joint Computer Conferences. Don't have it all common. Set up the Joint Computer Conferences separate as a corporation with a known distribution of the funds and known sponsors of that. Then AFIPS can take in all the ones that it wants to represent them internationally." And Paul did that.
One of my best advisors was also from Rand. What was his name? He did all the legwork for these conferences for years. One thing he suggested was AFIPS is representing these different societies; you ought to have a meeting where you get together with the principals of these societies and talk about the program and what's being done.
Aspray:
Did AFIPS at any time have the role within the United States as spokesman for the information processing community?
Harder:
Not really officially, although the placing of the office in Washington was certainly with that in mind. They were right there where they could be in on all of the legislation going on and everything.
Aspray:
And what about AFIPS having its own technical committees and such. We know that there were strong ACM committees, and there were strong IEEE committees. Were there strong AFIPS committees also looking at technical matters?
Harder:
I don't think so. I think this Education Committee that Don Thompson headed was a special thing that really wasn't much of a committee. He was just education chairman. We didn't have committees looking at any of the technical aspects of computers. That was the role of the societies, definitely. We were not encroaching on that at all. Representing the United States internationally and running the Joint Computer Conferences was definitely our role. To me if there are no more computer conferences in the United States, AFIPS role is practically gone. Somebody wrote recently making it clear that it still has that function of representing it internationally. It's just as strong in IFIP as ever. In the United States, support of IFIP is as strong as ever. That's about the only thing that AFIPS now does is represent the United States internationally.
Other Professional Societies
Aspray:
You held office in IFIP also, didn't you?
Harder:
I was the treasurer for three years. Actually, I was the treasurer for six years, but there wasn't any treasurer position. When I got into IFIP, they made me chairman of the Finance Committee, where I did exactly the same thing I did the next three years as treasurer when they changed the organization so there was a treasurer.
Aspray:
So this was in the early to mid-'sixties?
Harder:
Yes. I was president of AFIPS through 1965, and I guess in '66 about the next six years I was involved in IFIP. I finally resigned in '72. There's an article in the book by Heinz Zemanek, the IFIP Anniversary Book, about the financing of IFIP. Yes, I wrote the article. Heinz was editor.
Aspray:
What other professional societies have you been actively involved in?
Harder:
I was involved in power systems, but I was never an officer in it. I presented a couple of papers years and years ago. I wrote one on "Principles and Practices of Protective Relaying in the United States." Fig. 2 that the covered principles used in the various relays had 26 parts, A to Z. I presented it down in Mexico City first, and then essentially the same thing to CIGRE. Then when I was in charge of lightning field research, I felt that I had a general feel of what the whole field of lightning is. So I wrote about lightning protection for CIGRE, but never attended a meeting or anything. That's the National Society for Professional Engineers. I don't think there are any meetings. It's just something that professional engineers really should belong to. It establishes your responsibility as an engineer. Particularly if you're in the building trades or anything like that, you ought to belong to it. I got into it on the "grandfather clause" when it was first formed. Subsequently you had to pass examinations to get into it. But I helped the fellow that was working out the examinations originally. And it started back in the 'twenties.
There's this paper I wrote for and presented in Mexico. I met a fellow down in Williamsburg a few months ago, Stan Horowitz. He was a relay engineer all his life, and he worked for American Electric Power, which used to have headquarters in New York City. The president was Phil Sporn. Phil was a sharp engineer as well as a superb executive. He ran the system very well, and he was as good an engineer as he was administrator. Horowitz told me that when he first went to work for American Electric Power in New York as a relay engineer, Phil Sporn handed him this Harder-Marter paper and said, "Here, read this." If he knew what was in that paper, he had a pretty good idea of what power system relaying was. Now I knew all the ones that Westinghouse used, but I got a co-author from Duquesne Light because he would know the GE relays, too. We wanted to make it inclusive. We didn't want to leave anything out. So he joined me as a co-author on it. I used to gravitate naturally to such general papers. I'd get into a field, reach a point where I knew the whole field, and then I would want to write the whole thing up.
After I retired, I spent ten years writing a book on energy, and it was the same thing. There are nine sources of energy, and I worked on it until I understood all nine of them. I visited them all over the United States and all over the world. In 1982 this book was published. It should have had that title, "The Nine Sources of Energy" because that's what it's about. But I had a title "Energy for All." And the editor said, "Oh, that isn't a good title. You want to make it something like 'Principles of Energy Production' or something like that." I went along with it, and that's what it's called. But it's a very poor title. It doesn't say anything...." Whereas if I had labeled it "The Nine Sources of Energy," if anybody asked me what it's about, I'd tell them it's about the nine sources of energy.
Computer and Analytical Departments
Harder:
The department grew up around me, but first I was made consulting transmission engineer in charge of lightning field research and then given the responsibility of building the Anacom. I guess I first got one or two people and then hired some more. I eventually had, 120 or 130 professional people. But it just grew like Topsy, and then the company decided there ought to be the Advanced Systems Engineering and Analytical Department because we were really doing a lot of advanced systems engineering as well. And at that time Monteith was in charge of the Apparatus Division, which was over half of the company. It was the concept that, well, the computer is on top, and then it branched out, and you had all this different apparatus. The company was to be organized that way. You sold the whole system, but the top was the intelligence part of it. So we made a study of control and found out that there were two $600 million businesses: one in control of motors and generators and electrical things and the other was in process control. Like the control of paper mills. It's a continuous process. They were each $600 million businesses, and we were only in one. We were supposed to be in the control field, so we employed Stanford Research to help us look at all of the companies in the process control field that we might acquire.
We ended up with Hagen Chemical and Controls out between the airport and Pittsburgh. The board went along with us, and we bought Hagen Controls and set up a Westinghouse division called Westinghouse Hagen Control Division; we bought the name, too. Their principal product was Calgon; it's used in dishwashers. They had had Mellon Research develop this for them. So they took the name Calgon; it's now the Calgon Company. And we used the name Hagen. Now it's Westinghouse Process Control Division; we've dropped the name Hagen. But we used it for quite a few years. Up 'til then, the Analytical Department was really a service organization doing computations for other divisions. This gave it a certain responsibility in the systems field to see that the company was in the field correctly represented. It was a good thing; of all of the divisions that have been sold, this Process Control Division is still in existence over here, and that's quite a few years ago. So first we had the Anacom, and then as Monteith moved up in the company, he decided we should run the AC calculating board, too. And so we had the AC board and the DC board and the Anacom. Then we got a Prodac computer, and we got a Pace computer, and the CPC and the 650 and later the 704. The 704 was the last of the vacuum tube computers that we had. It cost $30,000 a month to rent it. It used 70 tons of air-conditioning to keep it cool. You had to supercool it to wring the water out and then reheat it so that the air you blew through the computer didn't have much water in that would condense on all the wiring of these things. That was a room almost as big as the computer room just for the air-conditioning. And then the very next one had transistors that didn't require any air-conditioning.
Aspray:
What was your first transistorized machine? Was it an IBM?
Harder:
Yes. It was a Seventy-ninety. Then a 7094. Then we had a couple of CRAY machines — not in my department, but in Westinghouse. One is a supercomputer for Pittsburgh, and one is ours. At that time we had the 7094, but the computer load just continued to build up. In spite of the fact that all of the other divisions got computers, the load here continued. When they got a computer in the Nuclear Center, they had to go to a CRAY computer with a big enough capacity. There just seemed to be no end to it.
Aspray:
I don't know much about the kinds of applications there would be. Could you give me a list of some typical kinds of problems that would be assigned to your group?
Harder:
We'd had these small machines. We worked on the blade problems, and nozzle problems with nozzles impinging on blades and the frequencies involved there. There were always the electrical transients problems. We had a lot of those. And lightning — how it works in power systems. Then in mechanical transients problems there are oscillations that occur when you start machines. There different kinds of drives, so they all had to be investigated from the mechanical transients. Are they strong enough? Then you start them and you get shock excitation, and you build them up through various speeds. So these were the earliest problems. You might classify them as saying these are problems that exist in design.
Well, the next whole series was design. I had fellows who would move to my department bodily for two years or so to write the program for designing an induction motor from scratch, making all the decisions, all the optimizing and everything. There's a lot to the design of a large induction motor. Guereny Godwin and Duchor moved into my department. Of course their department bought the computer time and paid for their housekeeping and all. Then a couple of fellows came to design turbine generators. Well, turbine generators in large systems are just entirely too big to design by computer. They should all be given separate consideration they're so big. So the manager of the Large Rotating Machinery Department said that that program for designing large turbine generators was well worthwhile for making parametric studies alone. Before they couldn't afford to design a whole line of machines to see what happens when you change the end-ring length or change this feature or change that feature.
But with a program, you can, and you can run parametric studies. He said that it was very valuable for that, but if they get an order for a turbine generator, they're not going to design it with that computer. They have a lot of information in the department, and a lot of things that have been tried and that a customer wants in features and so on and so forth. The transformers at Sharon worked out the same. They designed those by computer. A lot of the aircraft things at Lima were designed by computer. Even switch gear where it's mostly specifications rather than drawings. So there was a whole period where design was building up. Now smaller parts that companies stock, they don't design every order; they design it and stock it, and then you sell it from stock. Then a telecomputing center is used then to store all of the available stock and all of the customer information. When a salesman enters an order, it goes in there. It draws this out of stock, and makes out all the bills, and gives the Treasury Department all the information it needs, and does everything. The salesman in the district can practically initiate it, and it goes through automatically. But a lot of the things built in East Pittsburgh and Sharon and South Philadelphia, Lima, other plants, each order was designed. So that was done on the computer.
I think I told you that when we got the order for the Polaris missile launcher, we did all the calculations in my department. The skills of the engineers in my department: they could handle a coordinate within a coordinate, moving with respect to each other, and the equations and what this did to the dynamics of the system, and handle the whole thing and get it right. For several years after those Polaris missiles were in operation, nobody had made the calculations but us. Later other contractors did, and ours were all right. But, as I say, I could never have trusted the people when I first had them because they didn't quite understand that you can't just assume that calculations always come out right. Now I was under this impression: when I was a younger engineer, I worked on these railroad electrifications. We calculated them, we put them in, we never checked anything; and they all worked. I got the impression that anything that I calculated would work the way I calculated it. The first big blow came in that HCB relay out around Terre Haute. It tripped for external faults, and it wasn't supposed to. Well, what had happened I had the data on telephone cable. I knew the capacity of was 0 .08 mics per mile. But at Terra Haute they used a super light, thin wall rubber cable for the pilot wires up there. I looked it up in the handbook, and here paper and rubber have the same SIC. So I used the same constant for this cable.
Well, when we came to test it, it turned out those cables had five times as much capacity. I got to thinking about it later. The telephone cables aren't paper-insulated. The paper's crinkled. It just separates the wires. It's air-insulated really. This rubber wasn't really rubber, either. It was some compound that they use. There was a 5 to 1 difference. Well, fortunately this occurred right at the beginning. At that time we were using a system that for a through fault it balanced the two voltages. No current flowed through the pilot wires. But unfortunately the charging current did flow, and it was this higher charging current that caused the misoperation. We changed to a system where for a through fault you circulate current; charging current doesn't matter. Then when they have an internal fault, the charging current just adds to the tripping current. It doesn't hurt anything. We caught that after we had only about three relays out, and so we had no orphans. Well, it was just luck. I mean, that might have happened way, way into our development. It gave me a scare; it was the first time that I calculated something that it didn't work out like I calculated. It was a lesson. stock and all of the customer information. When a salesman enters an order, it goes in there. It draws this out of stock, and makes out all the bills, and gives the Treasury Department all the information it needs, and does everything. The salesman in the district can practically initiate it, and it goes through automatically. But a lot of the things built in East Pittsburgh and Sharon and South Philadelphia, Lima, other plants, each order was designed. So that was done on the computer.
Aspray:
What level of mathematics was used to solve these problems you've been telling me about?
Harder:
Many involved differential equations, some part arithmetic.
Aspray:
What was the training of the people you had on your staff?
Harder:
They were mostly electrical engineers. There was at least one physicist. Later there were mathematicians, hired after we got into digital computers. But at first in analog work, they had to be engineers; a mathematician couldn't have worked in the analog field. He wouldn't know the equations of the electrical circuits. We even came to a stage where we could hire girls for programming. That would be unheard of in the early days of analog where they had to be good engineers to do anything with it. But when the transistor first came out, I sent Jim Carleton down to Bell to pick up the know-how on transistors. We hired a young fellow, Bill Rowe, and put him to work on development. Before long, he came up with what was a NOR circuit and with the astounding knowledge that if you had a NOR circuit, you didn't need anything else. With the NOR circuit you could do all the logic that you could do with anything. So he made a little model of a NOR circuit. But were not a manufacturing department. How were we going to use this important information?
So I went up to Buffalo to the Control Department where they used logic in all these control systems and I told them, "Well, you ought to hire Bill Rowe and bring him up here because he's the one that invented this thing and designed it and knows how to use it." The manager up in Buffalo said, "Well, what are you doing telling us who we should hire?" And I said, "You don't have to hire him if you don't want to. I'm just telling you that he's the guy that can put this thing in operation." Bill moved to Buffalo, and before he got through he had built millions of those NOR circuits. I don't remember what superseded that logic; but they built millions of those models up there. He did a great job. Then he left the company, and he's with MITRE Corporation now, down in Washington. The last time I ran into him he was pushing this sun's illumination in the Sahara Desert and transmitting it to other countries by satellite. It was crazy as far as I was concerned, but then he was that kind of a guy. He'd push anything. The NOR circuit was a good sound development. He was a physicist, and he had worked out his master's degree on this atom-smasher.
When we got into inverting large matrices and things like that, they had a Dr. Long who came from W&J. He had an older brother who is a mathematician, too. At that time that he was going through school, W&J was specializing in mathematics, and they had some of the best teachers at W&J. So I hired him, and he's now out in Phoenix. Of course he retired many years ago, but he's still working with the power company on their computer work, with things like integrated circuits with chips.
Aspray:
What were your responsibilities during this time?
Harder:
Well, when the group got to a certain size, I formed sections and section leaders that handled everything under me. One group ran the computers, one group handled a lot of the analytical work and machinery and things; one group handled the problem of building up computers in other divisions of the company. It was well-enough subdivided that I could be away for several weeks over in Europe. I was in Europe 30 times over my career. My wife was along on 20 of those trips. Often they would be two or three weeks, and I'd come back and things were running just as well as before I left. I had a very capable secretary, but after 20 years, I felt I could run the Analytical Department with my hands tied behind my back, so I decided to retire at 60. I was interested in getting a broader idea of the field of science, as I told you. Yet, I never really was retired, because Monteith called me downtown and induced me to continue on as a senior consultant, which I did half time for another five years.
Rebuilding Siemens
Aspray:
And what were your responsibilities when you were a senior consultant?
Harder:
Well, Westinghouse had a very broad license with Siemens in Germany. We could interchange almost any kind of technical information, but you have to work at it to do it. So my boss wanted me to go back and forth every other month to Siemens to visit all their plants and laboratories and find out what they were doing at Siemens in Germany and find out how we could use the information here. Well, I was only working half time. I would go over there and do some things and come back and do some things with what I learned. I did get over five times one year, but mostly it was more like three times for the five years that I was working. So I was getting more and more of the Westinghouse engineers to visit Siemens. I would uncover things that were of interest. I remember in the nuclear field, I found out that Siemens had a way of putting boron into and out of the water by an equilibrium process. At one temperature and pressure it would dissolve in the water, and at another temperature and pressure would go out of the water.
I got a consulting engineer from Nuclear to go along over with me. He was taking notes galore. It turned out that the boron in the water is a shim. You have rods that control the reactor, but to get the starting point, the boron is a shim to where you want to be operating. Siemens's reactors were all baseload. They had no need to change. But they were developing this in the laboratory. They understood it and were developing it. Apparently our fellows had never heard about it. So I took him over, and we put it right into use in our reactor, and it saved two or three hundred thousand dollars over a job. To get the density of boron down, we didn't have any way of taking it out of the water and putting it back in. So that was probably the most spectacular of the things I learned. But I learned an awful lot and made a lot of friends over there. There were a number that came over here, too and visited our plants. Particularly after the war when Siemens was all shot to pieces, and they had to build up almost from scratch. They got an awful lot from Westinghouse: how to build circuit breakers and all the standard electrical equipment came from here. We learned quite a lot from them. So that was my main job going back and forth to Siemens.
But there were a lot of others. The vice president of Engineering that I worked for was responsible for watching all of the divisions, their engineering, and helping them with things. I got into a lot of that. Well, the company changed after that. Instead of having a top staff, they changed to six separate companies, and they no longer had a vice president of Engineering. The organization was different. While the present president, he was executive vice president, he cut down to the only services you got from headquarters were those you wanted to pay for. You wanted it bad enough, you were willing to pay for them. Headquarters didn't do anything that you didn't absolutely want. Now he's chairman of the board. He practically doesn't have a top staff. He doesn't even have a president. He operates with two advisory committees. One is called MAC, which means Marketing Advisory Committee; it's all the marketing people. They decide the things that a top staff would normally decide, but they are active runners of parts of the company. The other is EMAC, and that's Engineering and Manufacturing Advisory Board. So he has these two advisory committees, but he doesn't have a big top staff. I think he's in accord with the world handling of big companies because as I understand it, the way these big international companies operate, they don't operate with a big staff either.
They cut it all down to a bare minimum and depend on the people in the divisions to come up with programs and so on. This is all since my time. But unfortunately at the end of the five years that I spent increasing the use of this broad interchange program between Siemens and Westinghouse, they cut it all out. Everybody was going to bigger equipment: Bigger generators, bigger transformers. They built bigger railroad cars in Germany, and the German government decided that with the Common Market in Europe, they didn't need to have two big companies in Germany. They had plenty of competition from Switzerland and England and France. They didn't need to have Siemens and AEG. So they permitted them to go together and form joint venture companies to build all the big equipment. GE was connected with AEG, and Westinghouse with Siemens. And we were already operating on a consent decree for anti-trust. And our lawyer said it's impossible. You just have to discontinue the whole relationship. So after I'd worked on it for five years, the whole thing was dropped, and we discontinued all the amicable relations with Siemens. We had to just discontinue this interchange license because it would just have been impossible to have engineers visiting these joint venture companies in Europe and not have a flow of information from GE to Westinghouse.
Growth of Involvement in Computation
Aspray:
How you got involved so thoroughly in computing or calculating in your work? That wasn't a precondition when you went to work for Westinghouse.
Harder:
No. When I went in General Engineering, my job was on railroad electrification, and it was all calculating. So when they made the AC calculating board, I was familiar with the sizes of reactances and resistors that they would have to have in this machine. So I helped a little bit with the design of the AC Network Calculator, as it came to be called. Then for eight years I was assigned to the Middle Atlantic District, and that included all of World War II. After that I became consulting transmission engineer and largely took over McCann's work. McCann went back to Cal Tech to become a professor. He had had this concept of combining all these special-purpose analog computers into a general-purpose machine. Since he left, why, that fell to me. So I, in effect, was assigned the job of consulting transmission engineer, which he had held. I had the job of guiding the lightning field research, which he had been doing. And he had just started this project of building a pilot model to see if you could put all these things together. By 1946 the pilot model was done. Then he left in '46, and I just fell into the computer work from there on. As soon as the Anacom was built, it became the main computing device in Westinghouse. All the divisions used it. We had problems galore. This was all before the days of digital computers. But from the end of the war, digital computers were being developed and built. Other people were involved in digital so in the IEEE it got divided. There was a digital subcommittee and an analog subcommittee all in the Computer Committee. My part with it was analog at first. And as Westinghouse never really did build digital computers, why, I remained a user all my life.
Aspray:
When you joined General Engineering, did you know that it was mainly calculation that you'd be doing?
Harder:
No, I didn't. I knew they worked on systems, not on the design of a particular piece of equipment. That appealed to me. The fact that there was going to be a lot of calculation, I might have gone into that same department and been the inside man for one of the field men and just been handling his work.
Aspray:
Obviously the work agreed with you. You could have at some point later in your Westinghouse career left this work and done other work in the company, I assume.
Harder:
Oh, yes. Well, during the Depression that department was broken up. The only ones left were the ones that visited the districts. And I went to Switch Gear from 1933 to '38, for five years. That was not General Engineering. I was working on the Pennsylvania Railroad electrification. In that respect I did the same work in Switch Gear that I was doing in General Engineering. I carried it over, did it under their auspices. Then I designed some of the equipment for Hoover Dam Power Plant. That's design work; that's not general engineering. Then I designed voltage regulators for the Safe Harbor Water Power Corporation.
Aspray:
After the war did you consider moving off to another area and not being so intimately involved with calculations?
Harder:
I had gone back to General Engineering, and I had the Middle Atlantic District for eight years, from 1938 to '46. Then after that I was made consulting transmission engineer, and this was a step up. This was a nice job. I had a lot of interesting things to do. I had a lightning station on top of the Cathedral of Learning down in Oakland, and we had equipment on the fire towers all around Pennsylvania and West Virginia and Maryland. We had klydonographs that were devices that measured lightning: rotating wheels with a lot of magnetic strips round the periphery. As the strips came by, they gave you a time record of the current in the lightning stroke. And on top of the cathedral we actually had a cathode ray oscillograph. We couldn't afford that at all these other stations, but we did at this one station. So this was a lot of interesting work, and I was making a lot of talks, too. I had the freedom. I spent five weeks on a trip to Mexico City with my family. Some work along the way, some talks, some presentations in Mexico City. My six-week trip to the West Coast, similarly, was talks and visits as a consulting transmission engineer. So I couldn't have asked for anything that I liked better. This was just great.
Aspray:
Was there any responsibility within the Analytical group for computing outside of power? For that matter, did Westinghouse have business outside of power at this time?
Harder:
Well, Lima was aircraft electrification. By power, if you include industry, like motors, like Buffalo Motor Plant, they supplied the motors and control for all kinds of industry. Now they're power users. They're not electric utilities, but they're the users of power. The same with steel mills and paper mills and all of industry. Westinghouse dealt with all industry.
Aspray:
Westinghouse was building control devices, for example?
Harder:
Yes.
Aspray:
Did you do calculations on those as well as on power utilities?
Harder:
Yes. A lot of my inventions would have to do with control. A lot of them were on relaying, which is mainly utilities but for the transmission lines and equipment of electric utilities. A lot of the control devices were for industry, too.
Westinghouse in Depression and War
Aspray:
What kind of an employer was Westinghouse?
Harder:
Well, when I signed up with Westinghouse, it was just competitive. GE and Westinghouse both came to Cornell recruiting. So did Bell Labs. They all offered the same wage. There was no price competition in starting: it was on the order of $100 a month. Then the Depression came along a few years later, and so many people were laid off that didn't have any jobs at all and had to go home and live with their parents or scrounge around, that you were so lucky to have a job. I was so lucky I was working on the railroad electrification, that they went right ahead with it. That it never would have occurred to me. We had some furlough days, and I went downtown to see if I could use some of my great talents with some of the industries down there. I was out with an industrialist looking into this black plant, and he was almost crying. He didn't have anything to do. Didn't have any work for his plant. He'd had a going concern, and he showed me inside this building. It was pathetic, absolutely. And so I was so lucky to have a job that I liked as well that I would never have thought of leaving. There was nothing to go to anyway. Then when the Depression was over, as I say, then came the eight years in the district in Washington with the Navy.
Aspray:
Did the company try to rehire people they'd had to lay off during the Depression?
Harder:
Yes. They rehired a lot of them. So the same ones came back. Some of them went to different plants. In General Engineering there was a more experienced man than I working on the railroad electrification. But he was making $100 a month more than I was. Well, the boss could save $100 more letting him off. I could do the same work. He came back to Sharon instead of Pittsburgh. But a lot of them came back.
Aspray:
What about during the Second World War? Of the number of people [electrical interference sound].
Harder:
Whereas Westinghouse had been taking on, like, 300 students each year, there were none available.
Aspray:
In terms of the company, would the company agree to hire back or were they guaranteed that they'd hire back people who were called up for the war effort?
Harder:
I don't think so. In other words, they were separated from the company, and it would just be a matter of chance. But one thing, I was teaching. I taught many different things, but circuit theory a lot. After the war we had the selection of men that had graduated for the previous four years. Well, here I had class; they complained about the work I gave them. But they all did every problem, and they got them all right. In the end I was going to have to give practically all A's. What in the world will the University of Pittsburgh say? So I sent them down all A's, two B's. A big class. I never heard a word from them. Some secretary entered it but it never came to public notice. But it would have been unfair to these fellows. If had they come in the four years during the war, they would have been the top men in each class. Well, here they were all at once, and they were motivated. They had lost four years of their lives in the war. They were anxious to get ahead. They were married, some of them. They were by far the best class I ever had.
Aspray:
Were there any other companies that were considered direct competitors to Westinghouse besides GE?
Harder:
Allis-Chalmers was the electrical company. Cutner-Hammer was in motors and control. Of course our salesmen had to compete with them. But I never had any direct connection with that. Really, after I got into the Analytical Department, it was almost an internal service organization. I was handling the problems with the various divisions and teaching them how to do their work on computers, getting computers into the different departments. So the competitive aspects dropped out.
I established a costing area, and we had no trouble at all making out on our own initiative. Monteith should have done that a few years before; but we had a big surplus each year. For example, Gulf Oil in Pittsburgh, decided to go into the computer field. Well, rather than go to New York and use IBM Service Center, they'd come over here at noon, over to Westinghouse, and use our computer. Much simpler. They paid us the same amount they would've paid. I had a $200,000 surplus — just unexpected income from letting Gulf use our computer. So financially it was easy going. Never a problem. The only real problem, as I told you before, was when the Anacom was built. I hadn't been the one that had ordered the equipment or estimated it or anything, and then it ran way over what they expected. That was an embarrassing situation. But that passed over. We got over that hurdle.
Aspray:
It's paid for itself many times over.
Harder:
Oh, many times. I'm sure when it was built nobody would ever have believed that. Even though they didn't know about digital computers then, they would never have believed that 42 years later that machine would still be working. With the same internals that it had 42 years ago. The same resistors, the same reactors, transformers, inductors, low-loss devices.
Work Atmosphere & Organization
Aspray:
Let's talk about what the historians call "corporate culture." That is, what was it like in the workplace?
Harder:
I was looking at a tough situation during the Depression: whether I was going to get a pay, and there were deductions for various things. There were benefits. There was some employee life insurance. All my four sons were born in the hospital, and I know Westinghouse paid for all of that. I never had a coffee break in my whole career. I never wanted one. And in those days I and the people I worked with were all too busy, too interested in what they were doing, to be the least concerned about the fact that they were working all morning — longer hours than they do now. The work week got shortened. We used to work all day Saturdays to begin with. Then we got a half-day on Saturday. But Bob Evans and Charlie Wagner came in two or three nights a week, worked on this book, Symmetrical Components, which is the bible on that subject. I have 150 technical papers. A lot of them I came in nights and worked. I remember even during the Depression when I was transferred over to Switch Gear, I would come in nights. I had laboratory facilities that were available to me that were just like a kid going into a toy room. I just couldn't stay away from it.
Aspray:
And would you find lots of other people there when you went in the evening?
Harder:
No, I wouldn't. But to me this was an unheard of opportunity to use all this stuff. Here were power supplies. I darn near killed myself a couple times, but I was lucky. I got out of it.
Aspray:
What about educational benefits?
Harder:
At first the Engineering School provided 15 credits towards a masters degree. They offered the apprentices courses, and there was a training school. I got a job as one of the instructors. For two years I taught three mornings a week for two hours a morning. I taught the equivalent of a full year of college to these apprentices. They were simply high school graduates, but they were coil-winders. I gave them enough algebra and electrical engineering that they could understand what was going on in these machines that they were winding. Whether it ever did them any good or not, I don't know. Culturally it did because they got a lot of my education in the course of that two years' exposure. I don't know when the company started really encouraging back-to-school. It was certainly after the Depression. When Monteith became director of Education we used to have a big old building in Wilkinsburgh that was the Westinghouse Educational Center. Then he did so many other things. He had a vivid imagination, and he just moved from one thing to another and made the very best arrangements for Westinghouse employees to use the educational advantages that were available. Westinghouse would always pay half of the tuition, and they'd pay the rest of it if you went ahead and got a degree, got a certificate of some kind.
Aspray:
And you could go to any institution in the area?
Harder:
Anything in the area.
Aspray:
Did it have to be in an area that was demonstrably relevant to work?
Harder:
My secretary, Joan Pollitt, got a degree. When the company gave her the rest of the money, she threw a big party and invited us all to it. I don't remember particularly the courses that I took. I know I probably got it back. I don't even remember. I know the company paid half of the tuition at the time. I had to pay the other half. Education was certainly looked up to and was strongly supported. There are a lot of meetings in the company that are of an educational nature.
For example, they might have a lot of customers in and have lead engineers from the different departments come and spend a day discussing these different things. Not trying to sell them, just discussing what their characteristics were and how you looked at them. Well, today they have an annual advanced school that lasts three months. The customers have to pay now to come to that, but it's the same thing. They bring the lead engineers from all over, and some of it is credit material. For symmetrical components they get a credit at Penn State. A lot of it is excursions and tours through different parts of the shop and explanations of how electrical equipment works. My son took it last fall. One of their excursions was to the power plant down the Ohio River that has scrubbers. And I had never seen a power plant with scrubbers so I joined them and went down.
Aspray:
What about reporting structure and formality or informality at work?
Harder:
Well, there were always trip reports. You never went on a trip but what you made was a trip report. That was a requirement. If you spent any money going on a train — or later the airplanes — anywhere, when you got back a trip report was always required. And so that was formal. It must have existed before I came with the company.
Aspray:
And it lasted the whole time you were there?
Harder:
It lasted the whole time. It was always required, except when I was visiting the district regularly a week a month, I never made a trip report on that whole week. I had a lot of things that I picked up from the district that I had to do or have my assistants do at the office. But there my actual job was traveling. But until that, the trip report was a formality.
Aspray:
What about reporting progress on your work to your superiors? How was that done?
Harder:
Well, there was supposed to be at least an annual review with each man, appraising what he was doing and telling him. Then when we got into computers, we had a lot of people that belonged to the union. So I would be dealing with the union representative, discussing the progress of some of their members. That was always very amicable. On that I always had the feeling I'm dealing with people's lives now, and that what I do can affect their whole living and life. So I had to be especially careful. But the union representatives were all good friends of mine, and so it was never rough. It was always more than agreeable.
Aspray:
Was the company largely unionized, or was this something special in the computing area?
Harder:
All of the employees that were not in management were represented by the union. Started when General Engineering broke up and I went to Switch gear. They had a manager, K. C. Randall, and I think his principle of management was "you don't pay anybody any more than you absolutely have to; and if he will take it and work for you for less..." and so he was most unfair. Towards the end of my stay there, two fellows decided they were going to form a union. Tracy Weatherbe and another fellow who went out to Chicago. Well, they wanted me to join. I deferred. I was getting along fine myself. But as a result of that the 601 white-collar union was formed. Covered the whole company eventually. I told them, as long as you fellows are running it, I know there won't be any strikes, and it will be amicable. But if it gets in the hands of people that are radical, there will be strikes. And I was exactly right. That's what happened. There were long strikes a couple of times. Once when we bought the 704 computer — we bought it in '55 — and there was a strike all that fall and into '56. So IBM stored the thing someplace in Pittsburgh. Then we couldn't get the space ready. I mentioned, it took 70 tons of air-conditioning. Well up at Research at the R&D Center, the building was already equipped with facilities. So we could put the 704 up there for six months or so until we could get East Pittsburgh ready and then moved it down. But there were strikes. They were white-collar strikes. Several of them in the course of the years. But nothing before 1937. That was about the time that the union was formed. And for a few years after that it was all fine, too. Because it was in the hands of well meaning, responsible engineers who were simply trying to correct a very bad situation. We had a boss there, the same boss that gave me the 40 percent raise without batting an eye. He was the kind of a boss that made it necessary to have a union later. I won't mention his name.
Aspray:
In the postwar period, for you, where was your boss located?
Harder:
Well, first I reported to Harry Coleman who was right upstairs in the same building. He was manager of General Engineering. And then I was transferred to Engineering & Service, and that was Maynard Wyman, and he had an office right on the next floor. He was the most supportive boss. When he died, I just wondered who worked for whom. I think he worked harder for me than I did for him.
Aspray:
But what was the working relationship?
Harder:
There were staff meetings occasionally. There was always the Christmas party in his office. You weren't allowed to have alcohol in East Pittsburgh, but he violated that rule on Christmas. When I wanted to get the first IBM computer, it was the 701. I and Maynard went downtown and got approval for me to rent this huge machine that was going to cost $15,000 a month. Oh, we'd had the CPC and the 650 before that, but this was a big machine. He went downtown and went all through it. Then IBM came around and said, "Well, that 701 isn't so hot. You ought to get the 704, and that's going to cost $30,000 a month." Well, imagine going back to your boss. He never batted an eye. He went back downtown, he got approval, and when he got back he said, "I got approval for you. Get whatever damn machine you want that IBM is delivering at the time they deliver it." He said, "But I'm not going down again." His father was a sea captain in Nova Scotia. And he was a nice fellow, but when I came from Cornell all steeped in technology, a real hotshot young engineer, I didn't think anything of him at all because I judged people entirely by their technical skill. But he was the best manager I ever had. Later I learned that there are a lot of other things other than just the technology that are necessary to make things go in this world.
Aspray:
When he went to get approval for these rentals of the IBM equipment, did he ask you to prepare any kind of documents?
Harder:
No.
Aspray:
And you didn't go along and make any kind of presentation?
Harder:
No. I didn't go along. He just went downtown and got approval for me to buy it. I had the Analytical Department. I had other computers: the CPC and the 650. I think by that time the AC calculating board was in my department. But this was a big purchase. And interestingly enough, a lot of the load was going to come from Bettis. Sid Krasik was the physicist over at Bettis when it came through. We never had a written agreement at all. We trusted each other. It was all over the telephone that he would have so much work for me on this big machine, and we never put it in down in writing at all. And he was good for his word. As I think of doing business now, who would do that? Who would take a chance like that, that you were going to get all this work just on a fellow's say-so over the phone? He was a responsible fellow, and I trusted him, and he came through with everything he said he would.
Aspray:
I'm just a little surprised that there wasn't more of a selling job required to get the rental itself approved at headquarters.
Harder:
Would have been later. Later there was a group set up so you couldn't possibly get a computer without approval of this group. That came within two or three years after that. Sharon got a computer, and they had to go through this. Research wanted to get a computer, and their proposal came to me. And it was lousy. And I went to Sy Herwald, who was managing Research, and said, "If you really need a computer up here for window-dressing or something, I don't want to be the one to stop it," I said. "But these fellows are, in effect, saying that if you give them a computer free, they won't have to use that expensive one in East Pittsburgh." And Sy said, "I've never yet bought a computer for window-dressing, and I'm not going to start now." So it delayed them two or three years before they finally did get one. And that's mainly window-dressing because it was up there at Research where those fellows could use it, but we had one just a few blocks away down in East Pittsburgh. Research is where I pointed out the atom-smasher. And East Pittsburgh's just a few blocks down below. That's where my computers were.
Aspray:
When you got the 7090 for the department, did you have to go through a more formal procedure?
Harder:
Yes. The East Pittsburgh plant had gotten a UNIVAC for business purposes. I think both they and I knew that the next computer should really handle the engineering work and the business work. Well, the East Pittsburgh division was so anxious to handle it or to have it theirs that I said, okay. I had the major part of the work on it, but I let them get it and staff it with a group of technicians that actually ran it. And it never made a bit of difference to me because that wasn't my main forte. It was going to be $70,000 a month. And this was quite a step. The load on the 704 had been building up, but more recently it had rather slowly. So IBM agreed that we could have all the second and third shift free and put that money in a pot to help pay the starting-up of the 7090.
Something happened. The load started building up almost from that day that they made that agreement. By the time the 7090 came in, we had a huge fund built up. Then IBM came around and wanted to renegotiate this. And Tom New, who was handling it in East Pittsburgh, said, "Well, we both gambled, and you lost." So that's the way it stood. So we applied that huge fund to tide over the initial brunt. But the load built up so fast on the 7090 we were taking chances all the time. But in those days it always kept building up. And so in getting the UNIVAC, one of the main savings that they documented was East Pittsburgh had pay stations all over the place. Every little group had a paymaster. And there were about 30 of them. By eliminating these 30 separate little groups and putting them on the UNIVAC, they had practically paid for the UNIVAC in one fell swoop. But that savings was gone now these things were all on the UNIVAC. It all had to be reprogrammed for the 7090. I was running essentially a service organization, and so I was sort of stuck with IBM once all my clients had their programs in FORTRAN and IBM languages.
One time my boss came to me and said, Well, why don't you do like Bettis is and get this cheaper machine? And I said, "Well, if somebody gives me a million dollars to reprogram all of the program that my clients have on these IBM machines, I could do that, but...." He turned around. Nobody was going to give me a million dollars. So I was stuck with the IBM computers. But they did a good job, and my clients were all happy with them, and it made a good deal for a service operation like I was running.
Aspray:
For the employees in your service operation, what did you look for when you hired somebody?
Harder:
You had to understand the apparatus; you couldn't guess. So these fellows were not programmers. They were engineers, really. They were right in the heart of the engineering. Now when you get into digital computers, you can use some programmers, but it's still a mistake to not have the experts in the field. I mentioned that I at one time thought of going to China. You know, I was going to take as my topic "good and bad uses of computers." One of the bad uses was cases where programmers tried to program something for a division and the thing just didn't work that way. They didn't understand how a shop works. They imposed the structure. They assumed that each workman when he finished his day would be willing to sit down and write down what he did and put it on punch cards and you could keep complete track of everything. Those fellows weren't hired to do that. They would no more do that than fly. And so the complete program was a flop.
Aspray:
Did you actually learn that from experience?
Harder:
No. We never programmed anything but the engineering. But up in Buffalo they had a fellow who was just simply a programmer. He wasn't a shop man. And he should have gotten shop people in on his team, somebody that knew what shop people would do. Because they spent a lot of money on this program that was going to just keep track of everything. And always know where they stood on inventory and shipping schedule and everything. But instead it was a total flop. They couldn't use it because the shop people didn't work that way. That wasn't the way the shop worked. And so that was one of my examples of the poorer uses of computers.
Aspray:
Did you hire people who didn't have strong engineering backgrounds?
Harder:
Only after we had digital computers.
Aspray:
What skills did you look for in these people?
Harder:
Well, they were mathematicians. Although at one time we did hire one or two girls that were good at mathematics and did a very commendable job of routine programming.
Aspray:
Did you look for people who had already had some experience with programming to hire or did they just have math skills?
Harder:
No. This was too new. I don't think we could have found them. We had Dr. Zapher. He came to us. He was just a good mathematician. But it turned out that he was a fellow that anybody would trust. He was very trustworthy. And so actually he was a good manager. I had him managing the group that actually ran the computer. And he could deal with the union. He could make little deals — We'll do it this way, and these people come in and then next week these come in. Then I had a change of bosses, he didn't think that any mathematician could be a manager. So he immediately took Pete Zapher off of this job and made him consulting. Well, then the big nuclear center out there, they called me and asked what I thought of Zapher as a manager. I said, "I think he's great. He's one of the best I ever had." And they hired him, and he did a whale of a job of managing the big computers, the large computer operation they had there. He always knew the equipment. Then they took him downtown to head up computers for the whole company. But he was a mathematician. That was his training. He retired last year.
One job I sent him on: Westinghouse had a credit corporation that went into the finance business of financing the sale of their consumer products. And, of course, they wanted to make so much profit. And they had to have a schedule of rates that covered this, and they needed a mathematician to help them this all out. So I sent Pete Zapher downtown to help them. And he not only helped them, he found a wife down there.
Then I hired Dr. Wilson Long. He went to Washington and Jefferson, and then he moved out to Phoenix. He was helping the power company out there on their computing and particularly on big arrays of numbers. He's working on the mathematics is topology now. Working with a group that makes "chips." And he's working with companies that are trying to put all you can on these little chips. And he's providing the know-how in topology for that. But I remember on the 704 we were going to have a demonstration, and he programmed some long problem for the 704, and then we had a group sitting in the room there looking at it. And he reached over and pushed the button for it to start, and everybody expected to see lights flash and things and so on. Nothing happened! Well, after some investigation what had happened, the answer was already on the display. It had solved the problem while he was standing up. It was all over. We didn't have any concept how fast this thing was. We had the numbers, but it just really hadn't dawned on us how fast this thing was. So this long problem that he had programmed for the computer was all finished, and the answer was there on the board or the cards or whatever it had.
Company Hierarchy and Management
Aspray:
Can you give me some sense of where your various positions were within the company hierarchy?
Harder:
Well, there are three phases to a company like that: engineering, manufacturing and marketing. Then there are service operations within the company; there used to be quite a few of them. Well, there are laboratories, for one thing. There's a high-voltage laboratory, a high-power laboratory; there's the Analytical Department. They're all service operations that do things. Now ordinarily those are going operations, and when a manager dies and you'd have to get a new one, why, he is selected and hired for that. My case was entirely different. The Analytical Department grew up around me. I had one man at first, and I had maybe 140 at the end. And I was never appointed manager of anything. I was called a director, and then I was called a manager. But I was never selected out of any group of people that go and manage anything. But I managed the Analytical Department because it just grew up around me. So I really was not in the normal line of management selection for these things. Usually there's of course a lot of thought given to that, and the next higher management is always selecting the managers. Sometimes with pressure from special interests and so on that insist that this guy has to be the next manager. In my case there was never any competition. There was never any appointment made. The department grew up with me; and after I'd run it for 20 years to 1965, I figured that I wanted to get a broader idea of science, and it would be more interesting for me to retire and study science more broadly. Which I did.
Aspray:
Well, let me get some sense for where you were, though, within the company. You reported to a position with what title?
Harder:
They called it Industry Engineering after the Depression. Before the Depression it was called General Engineering. And it had a group on central stations. But after the Depression, when it was reconstituted, it was called Industry Engineering. And Harry Coleman was manager of Industry Engineering, and I reported to him.
Aspray:
And who did he report to?
Harder:
He reported to Monteith. At that time there was an electric utility division of the company and an industrial division. And they called the whole shootin' match the Apparatus Division, and it was more than half of the company. And it all reported to one man — Monteith. He reported to an executive vice president called George Wilcox. And the executive vice president was on the staff of the president.
And so when I wrote that memorial tribute for Monteith, I went downtown to see the fellow that headed up Personnel, the vice president of Personnel. And he said, "You must get a hold of George Wilcox. He was the vice president that Monty reported to during most of this period." And he said, "He's retired, but he still has an office here." So I got George Wilcox. And then we got Charlie Weaver, who had run the Bettis operation, first, to see what Monty 's activities in the nuclear field had been. And then we got Charlie Ruch, and Charlie had run the Westinghouse magazine all his life. Everything that happened got announced in this magazine. So he was a good historian. He could look up and see who was what when. He knew this magazine like the back of his hand. And so during a lot of this period Monty was running over half the company, called the Apparatus Division, which included all of Central Station, all of Industrial, all of Marine, all of Gas Turbines (we were building gas turbines at that time). And then the rest of the company was in other hands. Then I think, unbeknownst to Monty , higher up in the company they were planning a major change. And they divided the company up into six companies. They separated this Apparatus Division into parts. One was Central Station and one was Industrial. There were different vice presidents for each of these six companies. They did away with the headquarters vice president of Engineering. They didn't need him. They had six separate companies.
Aspray:
What happened to Monteith?
Harder:
Well, he ran the Electric Utility and Marine, Division for the two years of his career. And then he became senior vice president of the company for four more years until he retired.
Aspray:
Was the senior management of the company engineering or come from other fields?
Harder:
They came from other fields. Well, the president — at the time that we went into the nuclear field — the president was William Price who came up through the Mellon Bank system. One of the previous presidents had been George Booker, who had legal training. But when the decisions were being made to go into the nuclear field, William Price, this banker, was making the decisions with a committee of four. Lee Osborne was a senior vice president, and he was a businessman. He wasn't an engineer. George Booker was a former president of the company. Monteith was the only engineer of the four that knew Rickover well — and himself, William Price, that committee of four. When I got these together to write the memorial tribute for Monteith, Wilcox told me these things. And he was the executive vice president, so he knew. And Rook confirmed it from his records in the magazine. And so the whole decision to go into nuclear hinged on Monteith because he was the only engineer of the four. The only one that really knew what they were talking about. A banker wouldn't have the slightest idea. He was a good president. Westinghouse went through some hard times, and he got up in the auditorium and assured them that all this bunk that the business magazines were writing about Westinghouse — if they had one good profit report, they'd change their tune. If they had two, why, Westinghouse would be riding high and dry. He said, "They're that fickle." And he was right. I mean, it was a temporary hard time for Westinghouse, and he knew it was, and he rode through it.
Aspray:
Was there a sense that if one wanted to rise through the ranks in the company one should be in a particular area or choose a particular career path?
Harder:
I suppose there was, but I was not aware of it. I know the successor to Monteith in East Pittsburgh when he was rising up was Frank Benedict. And he was the one that I had to report to that the Anacom was over budget. Well, Frank Benedict was really put in there because he was a friend of some high, much higher, official in the company. And another group leader that thought certainly he would be the next one, why, he just left then. He got a job in the New York offices. He was flabbergasted that this fellow that hadn't even been in that line at all — he'd been in the laboratories — was picked out to become Monteith's successor there. Frank was a capable fellow, but he was also a darned good talker, you know. And he apparently had convinced some higher-ups that he was the one that should be picked for this. So there were things like that happened. But mostly it was the most capable guy that got in. Not always. There were some very bad things.
Aspray:
When you wanted to hire somebody new, how did you get approval?
Harder:
Well, you made up a budget for the year. And you had to forecast how many you were going to have to hire, and you put that in your budget. And you had the income to cover it and the people. I can remember one year that there was another of these 10 percent cuts. Well, I cut all these names out of the budget, but I hadn't even hired them yet. Got it down to where I was going to have to fire one technician, and he was the guy that maintained the Anacom. I wasn't going to fire any of my engineers, and so it was going to have to be him. But I had already let six names go out of the budget that I hadn't hired yet. But the money was in the budget for them. And so I went to my boss, and this meant Maynard Wyman, the one that I said was such a good boss. And I said, "If we let Art Ryden go, we don't have anybody to maintain the Anacom. We're going to have to use engineers to do this maintenance work on the Anacom." And he said, "Well, when they make these cross-the-board rulings downtown, they expect you to use all the ingenuity that you can. They expect that. They expect every department's going to use all the ingenuity they can to get around it, you know. So we'll put Art on our training budget." And for a couple of years Art was on his training budget. And he still was in my department, and he still did the same things he did, and eventually we took him back in our budget. And that's an illustration of what a supportive and good boss he was.
Patent Process at Westinghouse
Aspray:
Do you have anything else you want to tell me about management issues?
Harder:
Well, I think in the early days it had growing pains, as I mentioned, all starting in East Pittsburgh and then spreading out many miles apart to Newark and Sharon and South Philadelphia and so on. There wasn't the closeness, and they didn't institute the communications that there should have been. A patent on a relay could be made in East Pittsburgh, and the Relay Department didn't know anything about it. And there were similar loose connections that were growing pains that no longer exist. Now they've caught up with those things. And a few years later the Patent Department would never spend the money for a patent unless some department was going to use it and had approved their going ahead and spending that money for it. And then me being in a service department, the Analytical Department, I'd get into situations where there should have been commercial people in on it. The worst of them was that we built a real fine economic dispatch computer for West Penn Power. It was installed down here at Charleroi. It did a wonderful job for 30 or 40 years, and it should have been a product. It should have been sold all over the country. And I was too dumb to recognize that at the time or push it. It never occurred to me — I never even thought of it. All I was thinking about was getting that one built. And then I had other things to do. I was running the Analytical Department. But, as I think of it now, what an opportunity they missed to assign some group of people from the Analytical Department that did know what it was. And some salesmen put it in a van or something or a model of it and take it all around the country and sell it, sell it all over. Because it was a wonderful product that we just completely missed the boat on. Ten years later Digital Computers had taken over, and there was no longer an opportunity. But West Penn used this for 30 or 40 years because it had a display board and everything. And it showed how much power you should supply from each of the stations. And the dispatcher could handle it by telephone with the operators in all of the stations. So there were these failures — some due to growth. But in spite of that the company muddled through and did very well all through those years.
Aspray:
How did the company try to use patents?
Harder:
Well, mostly with GE it was a trade. You had a raft of patents, they had a raft of patents.
Aspray:
So you'd have cross-licensing agreements of some sort?
Harder:
There was a lot of that. And it was never used for the HCB relay. But patents were always valued highly, and when an engineer joined the company, he had to sign a patent agreement. Turn over all your patents to the company. To make it legal, there had to be legal tender passed between the attorney and the engineer. It was a new one dollar bill. Later, I think, pretty much at the end of my career, it got to be $25. And then there were a few special patent awards. I got a few of them.
Aspray:
So there was no plan to allow employees to share in the profits?
Harder:
Oh, no. No such plan at all. I think there have been cases where the company had no intentions of using that patent at all, they might release it to the employee if he wanted to see what he could do with it. But as long as he's an employee, he had signed a contract with the company to sign over all his patents to the company, and he just does that.
Aspray:
What about in terms of bonuses or awards or that sort of thing?
Harder:
Well, there were occasional special awards made.
Aspray:
But that wasn't really customary?
Harder:
I think I only got one or two out of the 66 patents. And that was completely unexpected. I think Newark just felt there ought to be a special award for that one. Recommended it. It went through. But it was no great increase in my income for the year. It was nice to get.
Aspray:
What would the process be for applying for a patent?
Harder:
Well, the engineers that had patentable ideas were supposed to disclose them. There were disclosure sheets, and you sent those into the Patent Department without any knowledge of your supervisor. It was just between you and the Patent Department. Then at the end of the month, they might have a little convocation in the boss's office, and all those that had turned in patent disclosures, might be stood up in a row and congratulated and awards, if any, given out. And representatives of the Patent Department would be there. And your management would be there. And it's kind of a big thing made of it.
And of course with 66 patents I was usually in all those pictures. If it wasn't for that, your co-workers wouldn't even know you were making patents. But they knew because I was almost always in this line-up in those years when I was working as an individual engineer. Later, as a manager, my patents were very infrequent.
Aspray:
It seems to me that that system wouldn't work well in that we all know that engineers are pretty busy, and it's a nuisance to do this paperwork in some sense.
Harder:
Well, you could say the same thing about technical papers. It's a nuisance. And yet there's a value to the company and a value to you as a professional for both the patents and the papers. And so I know when I went over to Switch Gear and Pete West could see that I could make inventions almost at will, he wondered if I shouldn't be coming in nights and doing this. I had that ability, and I wasn't making the most of it. Because I was turning in some patents, but I apparently in almost any field that I turned to I could see better ways of doing what they were doing. If I had that ability and I only worked eight hours a day at it and goofed off the rest of the time, I really wasn't making full use of my potential. But it certainly made life interesting to have these problems come up that seemed very difficult to solve and then come up with a unique solution to it that nobody else had thought about. It would occur to you, and that's a great feeling, you know. That's almost like winning a football game or something like that.
Aspray:
Did you see reluctance on the part of some of your colleague engineers to start the process of patenting?
Harder:
Well, some engineers are creative, and some just aren't. When we got separated, we had Relays in Newark, Instrument Transformers in Sharon, Carrier Current in Baltimore, General Engineering in East Pittsburgh. And we had what was called an Inter-Unit Committee, and I was generally chairman of it. And we'd get together periodically to discuss developments that required participation by all parties — instrument transformers were involved and the carrier was involved. And we might have these meetings once every month or every few months. And some of the fellows were creative, and there was one fellow that could think of more reasons why it couldn't be done. But I was just flabbergasted that a guy could think of so many reasons why things couldn't be done and yet never come up with things that could be done. And he was the extreme limit, but you had everything in between. Had some that were very creative. Some engineers in this group, in the Inter-Unit Committee, were very creative. And we carried on a lot of developments that involved all the divisions. That we decided on. That was the funny thing. We really had no authority. We were just a group of engineers who would get together, and we'd decide we were going to go ahead with a development. And he'd do that in his division, and he'd do that in his division. Being the chairman, I wrote the minutes. And I wrote what I wanted in the minutes.
Aspray:
Could you tell me what the process was by which a patent was prepared within Westinghouse?
Harder:
Well, you wrote out a disclosure and maybe a few claims and the patent attorney would take that. Maybe he had the authority to decide himself that that was a good patent. But he probably got permission from his boss to go ahead and write up a patent on it. Maybe just on that one disclosure, maybe, say, well, these two disclosures you can put together and make one patent. The engineer might not see it until he'd written up a case and was ready to submit it to Washington. And brings it around with a new dollar bill and a sheet for you to sign that —
Aspray:
So it wasn't necessarily an interactive process.
Harder:
No, not at all. No, the information you gave the Patent Department was supposed to be on the disclosure.
Aspray:
And were you ever contacted by the patent attorney during the process?
Harder:
Once or twice for some explanation of something. But I had an attorney by the name of O.B. Buchanan. And he was a real, real good attorney. And after writing several of my cases up, why, he got to know my field, really. So he didn't need much help. And most of the claims the attorney cooks up. I mean, you make the patent, but what are you going to claim that it does? Well, in the disclosure you claim a few things. But then he may come up with a dozen claims; they're written into the patent. And I think the value of the resulting patent depends a lot on the attorney. If he's good, why, he'll milk it for all there is to get out of it.
Aspray:
You told me that there was a change in patent policy about what to patent at some point.
Harder:
Oh, yes. It would've been in the early 'thirties that they didn't have this control. The Patent Department apparently could patent anything you wanted to patent. They made the decisions. But later — and how many years later I don't know — there was a company policy that some division had to sponsor each patent. And so they had to say, "Well, we're going to use it and will pay the costs of the patent. You can charge it to us. Before the patent attorney is allowed to spend his time writing up the patent and patent it."
Aspray:
Did that persist?
Harder:
Yeah. I think that's almost inevitable. It's certainly a logical way to run things. There may be some things missed.
Aspray:
Were you ever called in legal cases to provide expert testimony associated with patents?
Harder:
Not associated with patents. No.
Aspray:
But in other cases you were?
Harder:
In other cases. Other electrical phenomena.
Important Patents
Aspray:
I have in front of me this list of patents that you supplied to me. Could you go through this and tell me something about some of these patents?
Harder:
Well, this one that says "pilot wire relay schemes" occurring in 1933 — patented in '34 — was undoubtedly the HCB relay. Still in widespread use. A great value to the company. And a good example of a description of my patents, many of which were due to mathematical representation of the problem and then finding equipment that would do what the resulting equation said had to be done, had that as a very good example because that's exactly what that was. I knew the equations. The equations said what had to be done. And I found the apparatus that would do it and patented it, and it's still in operation today. So that is typical of my style of patenting. My style of inventing, I should say. They abandoned a lot of these. I haven't really looked. This out-of-synchronism relay — the Pennsylvania Railroad. Initially the electrification was just supplied from Philadelphia. But there was a supply down towards New York and then Safe Harbor down around Baltimore. If one or the other system slipped, there could be out-of-synchronism between those two supplies.
Aspray:
That was a patent for dealing with that problem?
Harder:
Yes. At Perryville there was a phase break, not just a trolley break: a place where there's a section of trolley that isn't connected to either end. When the panagraph glides first onto it and connects it to this end, then it's onto this dead piece and then onto the next trolley. So on the two sides if those systems are out of phase, it doesn't matter because they are never short-circuited together by the trolley. But the transmission lines are running all the way from Washington to New York, and they loop into these generating stations and go on. If two of those systems get out of phase at this phase break point at Perryville, you will find that the voltage dips when they go out of phase. If it dips three times within five seconds, you trip. That's what this relay was.
Aspray:
How have other companies handled that in the past?
Harder:
Other electrifications only had one power supply. This was the first time that the problem came up, and it was fairly obvious. We used it, but we didn't patent it. This vernier arm for a field rheostat we did patent, and we've used that quite a lot subsequently.
Aspray:
What was the problem there?
Harder:
The voltage-regulating system used on generators in those days was called exciter rheostatic. There was a rheostat in the field, and when you had to raise or lower the voltage to correct, relays closed that moved this field rheostat until you hit a button on the rheostat that created a voltage that was between the contacts of the regulator, you know. So then the regulator wouldn't try to either raise or lower because the voltage was correct. If you needed to go another voltage, why, again, this rheostat would get moved. Well, at Safe Harbor there was a waterwheel generator; there was also a frequency changer with a 60-cycle motor and a 25-cycle generator. They called it a frequency changer. Well, the frequency changer was rated for 100% overload for one minute. It would not only carry its load, but you could carry an enormous overload. The waterwheel generator you couldn't because the waterwheel wouldn't supply it. But the frequency changer you could.
Well, now, the field current required in a generator when it's supplying very light load is just a very tiny field current. When supplying 100% overload, it's a tremendous field current, and the steps on the rheostat are very small. And so this rheostat that controls the field has small resistance steps at one end and very large resistance steps at the other end. But the rheostat only has 200 steps. And you have to be able to divide up these resistances so that the regulator can always find one step that's right. Well, you could do that with ordinary generators that didn't have the 100% overload. But that extended the requirement so much — very small resistance steps at one end. And then to make matters worse, the transmission lines went through Baltimore on underground cables, and cables have capacity. And in the early days of electrification, there might not be many trains running. And most of the load was these cables in Baltimore.
And so the light load, the generators were almost self-excited and required extremely large resistance. And a 200-step rheostat just wouldn't do it. So you had to have a vernier. But are you going to make the vernier with great big steps for the low-load end, or very large steps for the other end? You would have to make a decision. So by using that vernier arm on the main rheostat — we used an additional arm; and it and the other arm spanned two steps — you would go beyond where you needed it and then back up. And you were short-circuiting these two steps first with six steps and then five steps and four steps and three steps and two steps and one step. And you got the effect of about a 5 to 1 increase in the number of steps on this rheostat. It was the equivalent to having 1,000-step rheostat instead of a 200-step rheostat. And that was my invention. That was the way I solved the problem, and they've used it ever since. They had many cases where this problem arose. And so that was one that came at the time of the Pennsy electrification. That was called a vernier arm for field rheostat.
The vernier arm rheostat was a much-used invention. Now during the Depression, the power companies were short of money. And the potential transformer for high-voltage bus, like a 132,000-volt bus, costs about $100,000. But you need it because you need that voltage for metering and relaying and for a lot of things. You need that voltage. Well, you can get it another way. You can use a low-voltage transformer on a 12,000-volt bus or low-voltage bus and then compensate for the drop through some power transformer with a compensator. They call it a metering compensator because it had to have metering accuracy. It couldn't just be a roughshod deal, but it had to be accurate. And that low-voltage pot would cost like $2,000, and the compensator, $6,000 — compared with $100,000 for the high-voltage pot transformer. And we were able to sell some of them. And we thought this was going on forever and patented it.
Well, after the Depression, as soon as the power companies got some money again, they forgot all about this. And they'd just buy a high-voltage — Actually, it's to their advantage. The Utility Commission agrees that you need a high-voltage pot transformer. They can put it in the cost schedule, and they make more money that way. It would be foolish for them to do all the engineering work to do this if the Utility Commission allows them to buy that pot transformer. Okay. But there's an interesting story about this. Northern Indiana Public Service bought one of the first ones. And they assigned their relay engineer, Erickson, to set it. He needed somebody to hold his hand while he was doing it, you know. He'd never done it before. So I went out to Indiana to help him set this compensator. And in the car on the way back to the train station we got talking, and he said, "You know we got this transmission line connected to American Gas & Electric. And I'd like awfully well to use Westinghouse relays on it. But they're going to have a one-cycle carrier on the other end, and I'm stuck. I got to buy GE." And he just hated like the dickens to buy GE; he liked Westinghouse. But he was going to have to buy GE because they had a one-cycle carrier, and American Gas was going to have one cycle on the other end. And he had to line up with them.
So I said, "Well, we have just discovered at Newark that we can use a single-phase directional element to control the carrier. And ours will be one-cycle, too." And that was all that was said. I didn't know but what he'd already bought it, you know. I didn't know. The next week in Newark we got a wire from Erdmann, the salesman in Hammond, Indiana. "Erickson understands we can now use the single-phase directional elements to control carrier. If so, we have an order." That was our first order for the new scheme that I told you about the other day. That was our first order. I was instrumental in getting it because I just happened to be out in Indiana talking with this guy about a metering compensator. That was just a pure coincidence. And it led to our first order. But there were very few of those compensators ever built. They certainly weren't worth the cost of patenting them because as soon as the utilities could get money again, why, they put in pot transformers and the heck with metering compensators. We spent a lot of money. We thought that was big stuff. They added a hot device we used. I did quite a lot of development work on the voltage regulators. It's pretty highly technical, and I won't go into that. But I marked it here because it was used and did solve problems. And it was used in the system from then on; it became a standard part of it.
Aspray:
Can you briefly explain what the problem was?
Harder:
Well, it had to do with driving this rheostat in the field. The system that I came up with was you closed a relay and the length of time it was closed depended on a capacitor. In other words, the current died out gradually. And so you could adjust it so that the rheostat would go one step. Period. Or two steps. Whatever you wanted it to. And with the previous system it would get all confused. I don't want to say more about it because it really was a complicated problem. Well, when I went over to Switch Gear when General Engineering broke up, we had the order for supervisory control of the Pennsylvania Railroad substations between Paoli near Philadelphia, and Harrisburg, to be controlled from Harrisburg. And the control cable went right along the trolley line, and of course, if you got a short circuit on the trolley, you got a high voltage induced in the control cable. And we were supposed to have supervisory control of these seven substations between Harrisburg and Paoli. And the only scheme that we had for dealing with this high voltage induced in the control cable was drainage. You just put resistors across to ground. And you could go one or two stations that way, but you certainly couldn't go seven stations. So Pete West gave me the problem.
And what I came up with was what I called the self-exciting neutralizing transformer. I don't see it among the patents here, but maybe it didn't get patented. I don't know. But it was an invention definitely. And the Telephone Company had a three-winding neutralizing transformer. The exciting-winding they would connect to the station ground and to some remote ground. And then if the wires here were going to be, like, 4,000 or 5,000 volts above the station ground, why, coming through this transformer they came right down. Because the voltage drop in those would be the same as in the exciting-winding. They're just one-to-one ratio transformers. They have the same voltage in them. Well, we accomplished the same result with a self-exciting, a two-winding neutralizing transformer. But where you had the transformer from the lines where it came to the station, you put a capacitor from each wire to ground. And this was like a half-mic capacitor. A one-mic capacitor has 2650 ohms. A half-mic has about 5,000 ohms. Well, that's at 60 cycles. At 25 cycles it would be two and a half times that. Maybe 10,000 ohms. But the exciting impedance with the transformer was well over 100,000 ohms. So in effect these capacitors grounded the whole business here to the station ground. So that the wires used on the switchboards were safe. And you used another one where you left that station and went on to the next station. And so we could handle the seven stations from Harrisburg using this self-exciting neutralizing transformer. And that was used — Well, it was first used there. That was my invention. I'm very much surprised not to see it. Maybe I missed it. Maybe it never got patented. I've always been citing that as one of my inventions. I certainly invented it, but I don't see it here. It may be that it never got patented. Earlier where I mentioned the HCB was a pilot wire relay all right, but it wasn't the HCB because here patented in 1939 is the HCB relay.
Bus protective scheme. Well, the reason there were problems with bus protection is that on a generator bus particularly, when you get a short circuit not far from the generator, there's a big direct current component that lasts maybe a third of a second — 20 cycles. And all that direct current going through transformers just simply saturates them. Well, if it was just one circuit, they would all saturate alike. But on a bus you have different amounts of current coming in different current transformers and some saturate and some don't. So some are correctly reproducing in their secondaries the current in their primaries. And some aren't. And so you get faulty indications that there was an internal fault when there wasn't. And so there were problems with tripping of generator buses. And a lot of things were tried. They tried harmonic restraint. They tried relays with a lot of different coils that were connected to different circuits. But I thought the best thing to do was to take the iron out of the transformer altogether and just use a linear coupler. And going back in history, when Westinghouse first wanted to test and calibrate current transformers, Chubb devised the idea of a perfect reactor. He took a ring of marble and graduated it exactly on a lathe. And wound a perfect toroid on it.
Now with a perfect toroid, if you put current in it, you get flux around through the marble inside of it; you get no flux outside at all. Because it's a perfect toroid, and any circuit outside of it isn't linked with it at all. But everything inside of it is a perfect linkage. And so they could calculate on paper what the mutual inductance was. So they knew exactly what the current was, and then they could use that to calibrate current transformers. Well, the Bureau of Standards picked it up. And that became the standard way of calibrating down at the Bureau of Standards. Then when I picked up the idea for using it for bus protection, it was like something coming home to Westinghouse that had gone to the Bureau of Standards and back to Westinghouse again.
There is in electrical engineering a principle of reciprocity. What a mutual reactance is if you have two circuits that are mutually coupled and a certain current in this circuit produces a certain voltage in that circuit, the same thing is true in reverse. The mutual inductance is a common property and has the same — and you just write it as a mutual reactance between these two circuits, and it works either way. The current here produces that much voltage in this circuit; a current here produces exactly the same voltage in this circuit. So here you have a perfect toroid. In one way you know how it works. You know that if you put current in that toroid, the flux is all inside that toroid. And all of that flux links any conductor that goes through the hole — no matter where it is. You can put it through sideways. And none of the flux links any conductor that doesn't go through the hole. No matter how close that conductor is outside, it gets no linkage to it at all. So you know that for anything through that hole — and we built these things to have .05 ohms at 60 cycles; that means 5 volts per 1,000 amperes — for every 1,000 amperes through it, you get 5 volts in this. Well, now, as I say, when you put the current through the toroid, you know that all the flux links this conductor through it. But if you put the current in that conductor, you don't have any idea that the flux around that conductor some of it cuts through the toroid and it goes all over the place. If you put the conductor outside the toroid, you might be surprised that there's no mutual inductance at all to that thing. But you know from the reverse that there isn't because if you put the current in the toroid that none of the flux links that conductor outside no matter where it is.
So there's a principle of reciprocity there. And we made use of that to make them astatic so that you could put these things in circuit breakers or like current transformers — anywhere — and as long as the conductor went through, you got 5 volts per 1,000 amperes. Period. And there was no saturation involved because there was no iron. So they didn't saturate. So that you would get perfect operation. And we built them into circuit breakers. We built them into transformer bushings. And we built them into separate devices. But we called it a linear coupler. And that's what this bus protective scheme is. I'll mark it linear coupler. And the first tests were down at York and Middletown on two high-voltage stations, and they worked exactly like we predicted. This friend of mine up at Sharon had an idea — Ed Wentz — he had an idea that since we didn't have any machine to wind toroids — you could have; you could have a shuttle that — but we didn't have any machine to wind toroids. They all had to be hand-wound. And they still are today. I talked with some of the fellows in these new operations, and they still have to wind toroids by hand. So Ed got the idea that, well, if I wind it as a solenoid on elastic material, I could bend it around then and stick it in a case and then just measure and see if it's elastic or not. Whether it works like it should. And I thought it had.
I always thought, all my life, that that worked. But when I got this award, Ed called me from out in Washington, and in the course of the conversation I recalled that, that he had wound these toroids that way. And he said, "Yeah, but it didn't work." I had always thought it had worked. Apparently it didn't work, and they had to wind them by hand. Same as everybody else did it. So I didn't learn that for 50 years later that that scheme hadn't worked. But anyway, the linear coupler was a perfect solution. If you're building a new station, everything is new and you can build these into the circuit breakers and all. But the fellows at this meeting in Williamsburg, Virginia told me that they're not so much used anymore because often the breakers are already there with current transformers in them. Or they want to have current transformers for some other use, too. And so they don't have the situation of a brand new station with a dedicated use of these things to this purpose. And so they put up with a little less perfect operation in order to have current transformers. So they're not as much used anymore, but they still are used. And this patent is '41. This is '91 — exactly 50 years. And of course the patent came a couple of years after the test. But the original tests with the York and Middletown stations down below Harrisburg, just down below Three-Mile Island, down there. I was going to mark that one "bus protective scheme linear coupler." About that time one of the relay engineers from Pennsylvania Power & Light in Allentown said, "Well, when anything new comes out of Westinghouse, why, Ed Harder has his name on it." [Laughter] Because he'd been familiar with all these other inventions, and so he just sort of expected by that time that I'd have my name on that one, too.
There's a lot of these that I would have to read the patent myself to know what it really was all about. So I'll just dwell on the ones that stick out. I see some names here — "Broad-Range Linear Rectifier." Well, with a rectifier, if you start at extremely low current and go to extremely high current, why, you get a ratio of ac to dc, but then it tails off at both ends. Wouldn't be linear for a long, long ways. And a lot of these things became necessary when we got into the Anacom. Needed a broad-range linear rectifier. Or we needed a minimum voltage selective circuit. You had a lot of voltages, and you wanted to get the minimum of them to do something. And so there was an invention of a device that would select the minimum of these. An arrangement of rectifiers, you know. The rectifier's biased in such a direction that it would only respond to the minimum of these voltages. And a lot of that was necessary because of working on the Anacom.
And the dates of these, we're getting into — Here, way over in 1943, I see "Air Core Coils for Bus Protection." Well, I've said this bus protection back here was linear coupler. I'll have to mark this one, too. I don't know which was which without reading them. Air core coils for bus protection linear coupler. But I always thought it was interesting that that thing started with Chubb Research, went down to Bureau of Standards, and I think that's where I kind of picked up the idea. I knew what they used to test current transformers, and they didn't have any iron in it so I worked it out for use as a bus protection scheme.
One of these is, you know, like a forcing function for producing sine and cosine decaying functions. That would undoubtedly be for a forcing function or an analog in the Anacom. See here — perfect transformers for an electric analog computer. I don't know what I did, but transformers have a lot of imperfections. And if you need a perfect one, why, I'd have to read the patent to see [Chuckling] what we did to make it more perfect. Often instead of a potential transformer, they used a potential device on power systems. And that's a capacitor coupling. Instead of a transformer, it's a capacitor coupling. And there were bushing potential devices that were built right into the bushings of circuit breakers and so on. And then there were separate coupling capacitors that were used for introducing carrier into the system or for getting the potential. And they generally weren't of metering accuracy, but they were quite adequate for coupling devices, for carrier or for relaying purposes. But I don't think they were really considered of metering accuracy.
But from my knowledge of circuits and equivalent circuits, I understood the equivalent circuit of the potential device. In fact, there was a principle that had been used in communication circuits quite a lot. It's called Thevenin's Theorem. Ever hear of it? Thevenin's Theorem. And it wasn't known in the power systems at all. And I ran across it in an instruction book to a device that we had bought from the Telephone Company that mentioned Thevenin's Theorem. And so I wrote a paper for the Electric Journal. I wrote an article on Thevenin's Theorem and how it worked. And how it works is that the equivalent circuit of a device has an internal voltage that's equal to the terminal voltage when there's no load. That's the internal voltage of the equivalent circuit. And it has an internal impedance that you would measure on these terminals if you short-circuited all the sources of voltage inside and measure back. That would be the internal — And those two — that internal voltage and that impedance — is equivalent circuit of that transformer. That's Thevenin's Theorem. And so when it came to studying these potential devices, the people who were working on them had no idea of this at all. And so I was able to work out a lot of things like that — practical applications of Thevenin's Theorem from really knowing what the equivalent circuit of this thing was. And then I could connect anything to it externally, and I could calculate how it would work.
There are some patents here relating to tandem steel mills. They came along at a stage when we were having all the troubles at Gary. Some of these developed into join patents. All these names: Stringer, Moore, Greenwood, Harder were all working together. The patents were joint on some of the things that we came up with.
Oh, I know. The Pennsylvania lines were single- phase. There were usually four of them along the right-of-way, two on each side of the railroad. And the single-phase transformers were grounded at the midpoint through a high resistor. So they were 132,000-volt lines, and that was 66,000 volts each side of ground. Now when electrification was being done, in order to detect a fault on the line, we built a relay that had two coils: one connected here, one connected here. And when one voltage collapsed, the other went up, and it was a differential voltage relay, you might say. It measured the fact that there is now a fault on that line because one has gone to zero and the other has gotten big. And those relays were not a very good design. A good design of relay, you start with a base and everything is fastened to that base and is rigid.
But this thing that was built out of one thing was added to another and added to another. And the Pennsylvania Railroad was having a heck of a time maintaining it. And some years later they requested Westinghouse to redesign that relay better — to make a better relay. The principle is all right, but make a better relay of it. And I took one look at it and by then I knew the equivalent circuit of these potential devices. And I said, "You didn't need to do that at all. You just connect these two wires together and use the over-voltage relay." So I went to Pete West and said, "Shall we tell them? Or should we do what they ask us?" And he said, "Well, we'd better tell them." So we just did away with those balanced voltage relays, which were hard to maintain and substituted bringing the two wires together from the two potential devices and run it through a single-voltage relay. Then if you needed time, you added a timer. If you had to give the next section time to trip out first, why, you just added a timer. But this was because by then I understood the equivalent circuit of a potential device, and I could see that you could easily do this. There was no patent involved. This was just one of the aftermaths of better understanding of how things worked.
See a lot of deferred things here. It's amazing that these things that were patented add up to 66. I've been clear through it, I think.
Aspray:
There are a few other things that am wondering about the patent process. How was it determined whose name would be the one that went on the patent? What was the underlying principle behind this?
Harder:
Well, mine was it was a team effort. And if two of us worked on it — no matter whether it was my idea — it was my idea we put both names on it. We both worked on it. And undoubtedly — Now some of those have several names on. That's where we were working on the trains and the head-end control of the trains clear through. And it was hard to say who contributed all the ideas that went into the eventual patent. But even if there was no question about it, if somebody else was working with me on that — Like that invention at Newark where I was loaned to them. It would have been foolish for me to just put my name on the patent. We all had the problem, and we were working together as a team. And while I came up with the idea, Goldie went into the lab and checked it all out. And so you'd find two or three names on the eventual patent. But there's no loss. You get credit for the patent whether there are other names on it or not. It's counted as one of my patents. And so this was just good personal relationships. Now I know some very bad situations in the company where an unusually greedy guy has really stolen things that other people had done. You don't have any friends there, I can tell you. I was going to invite this guy to meet you because there are other things that he's done that are interesting historically. But I wouldn't have invited these two together because the one has had cases — where this other guy has claimed things that weren't his.
Important Papers and Publications
Harder:
Peak voltage across saturating reactances is an interesting story. Shortly after I was in General Engineering, a big transformer blew up out of Louisville. A big tap-changing -under-load transformer. And the people at Sharon didn't know why it blew up. They had tested it in the test floor at full-load changing taps, and voltages were nowhere near enough to flash it over. And so there was a meeting down in East Pittsburgh, and they got all the top consulting engineers. And my boss, Bob Evans, took me along. And I was just a young kid out of school. But I knew what happened. I could see it. And when I explained my explanation of it to all these top consulting engineers, why, they let me handle it from then on [Aspray chuckling in background] with Monteith. Louisville was in the Chicago district, and it came under Monteith. And so Monty took me around to Chicago. Well, at first, I have to say that in addition to this one that blew up in Louisville, there were 18 similar ones all over the country that were built the same. And the company had to find out whether those were in danger and whether they might have a whole series of explosions all over. It would be disastrous!
So I did the analyzing.
Well, first let me tell you what the trouble was. If you have a circuit and an iron-core reactance in it, when the current is going through zero, that is not saturated. And it can have an extremely high impedance. Then as soon as the current has gone through zero and the flux is up to full flux in that iron, why, it's saturated, and the rest of the circuit it's as though it wasn't there. But even though it's just a little 5% device in a 100% circuit, there is a time when it's impedance may be high enough to soak up all the voltage in the circuit right across that little 5%. Well, now the way they changed taps, they had a little preventive auto-transformer that stepped along like this. And the tap came out here. And when it's a transformer — it's working like a transformer — when you open one contact, it's a series reactance. It's a saturating reactance. Well, now when the current goes through zero, it's going to soak up all the voltage available in the circuit. The way electric circuits work if a circuit consists only of voltages and inductances, the instantaneous distributions are the same as the steady-state. Now that wouldn't be true if it was R&X or other. But if the circuit has only inductances and voltages, then when the current's going through zero, if you write the differential equations, you can see that the voltage just divides proportional to all the inductances.
Well, the way they test transformers at Sharon — big power transformers like that — they put them back-to-back, and they circulate full-load current, and it only takes about 20% voltage because they only have 10% reactance in each one. And you can circulate full-load current. And they had circulated full-load current in these transformers, and they had measured all the voltages, and they weren't excessive. Sure this little thing was soaking up all the voltage there was, but there was only 20%. Now then, you get that same transformer out doing business, and let's say the power factor of the load is 80%. Then when the current is going through zero, there's 60% voltage. Not 20%, but 60%. Three times what they had tested. Or if this is a tie between two systems and maybe there's a 60 degree angle between them. Then there's a 100% voltage. When the current goes through zero in a reactance, why, the voltage is 100%. So you can get much higher voltages with these transformers in the field than they were getting when they tested them at Sharon.
And of course these are good transformer designers, but they weren't sharp young guys right out of college just like I was that could see beyond the surface a little bit. And so as soon as it was explained to them, why, of course they understood it. But then I worked with Monteith, and we were out in Chicago. And we analyzed all the transformers and the circuits around Chicago. And we agreed to change 12 of them, I think, and there were 12 more that I said were safe because they were located in a circuit where the voltage would not be over 20% when the current was going through zero. Either a high-power factor load-circuit or not very big angle between two stations or something like that. So I analyzed all that. Well then, I used that for my — I'm getting into some pretty deep theory here for — But so far it's simple circuit theory. But now these are three-phase circuits. And this changing of taps is only happening in one phase at a time so it's an unbalanced condition.
Now we have the method of symmetrical components, but the method of symmetrical components is only applicable for steady-state conditions not for transients. But, as I told you, if a circuit consists only of inductances and voltages, then the steady-state and the transient solutions are the same. Well now, since this phenomenon only occurred during a very small part of the cycle while the current was going through zero, why, during that part of the cycle I could assume that all the voltages across resistors stayed constant. Any voltages across capacitors stayed constant. Instead of treating them as IR drops and capacitor voltages, I treated them as emfs. Then the circuit fulfills the condition where it's only emfs and inductances and where the method of symmetrical components can be used. And so I don't think the people at Sharon ever did understand that. But it was necessary that somebody up there write the paper about the basic phenomenon because it had to be a transformer man, you know; they had to save face a little bit. But I wrote the three-phase story — peak voltages across saturating reactances. Well, Fahnoe and Maslin wrote the one for Sharon. And I wrote "Peak Voltages on Saturating Reactances in Three-Phase Circuits." And then I used that for my master's thesis also.
I have an interesting story about my Master's thesis. Pitt has a catalog out that tells the requirements of getting a master's degree or a doctor's degree. And it says plainly that the thesis has to be published. So I get this problem of DC in transformers. First time it's ever happened. Beautiful thesis project! All this was brand new information. Nobody knew anything about it before then, and so I wrote it up and published it. Took it down to Pitt and — "Oh, no! It doesn't mean that. We have to approve the thesis before you start working on it. You can't use that for a thesis." Why, I said, "Well, the book says it has to be published." "Well, it doesn't mean that. We have to approve it first." So then I came on this, Saturating problem and I did get it approved. They approved the fact that I could write the thesis on this subject. And I happened to be down in Professor Dyche's office one day, and I said, "Well, it's coming along fine. It's going to be published in the Electric Journal. It's going to be published in March." I was probably down there in November or December. And Dyche almost blew up. He said, "Well, you can't publish that." And I said, "Well, the book says it has to be published." He said, "Well, it doesn't mean that. What it means we publish it." And so we called Charlie Scarlott, who was the editor of the Electric Journal, and Charlie agreed to delay it for several months before he published it so that you could add a footnote that this is based on a thesis presented at the University of Pittsburgh. So he published it, after the thesis was accepted. But he just delayed it. He was a good friend, and he delayed it. So that's the story of how this finally became my master's thesis. [Laughter] But it also was a good subject because it was something that had never been run into before and had a lot of good theory connected with it and how you could use the method of symmetrical components to achieve the result.
There was a Pat McGee who worked in the railway section across the aisle from the transmission section. And we often worked together on the Pennsylvania Railroad things. And he was interested in generalizing some of this work. So he came up with fleets of trains — six-train fleets and ten-train fleets that you had to move over a railroad. And what sort of an electrification would you have to have for this? And generalizing it. "Power Supply for Main-Route Line Railways." And he and I wrote it up. And that became a paper. And during the recent energy crunch, the subject came up again of electrifying from Harrisburg on to Pittsburgh saving oil. You know, they were using diesel engines for all this. And they were using this paper down at Carnegie Mellon — fellows that were working on this proposed new electrification were using this paper. That was published in '33.
Thirty-seven years later they were using that paper. And then I told you about how I got into harmonics with telephone interference between power communication circuits. Well, here was a paper, "Effects of Rectifiers on Systems Wave Shape," that related to that. And writers on this subject use this later. This was recognized as an important contribution to symmetrical components. Then here in 1938 is the paper about that relay scheme I invented in Newark. And I see the paper is by Harder, Lenehan (Lenehan was a consulting engineer down there), Goldsborough was a real good relay engineer. And so when we came to write the paper, they were all involved, too, same as they are in the patent. And here's Thevenin's Theorem. As I say, I found it in an instruction manual of a device that we got from Bell Labs. Nobody in power systems knew anything about it. Never heard of Thevenin's Theorem. And so I wrote an article and popularized it. And here's paper #14on the HCB.
Then I have told you about the next one that the Pennsylvania Railroad — We ran a relay study for them, and I wrote the paper because they didn't want to have one of their men co-opt with Westinghouse and not with GE. So they agreed I should write it. That's the paper. It's the bible on the Pennsylvania Railroad relays. Then when I got into Switch Gear, that's where these potential devices were made. And Paul Langguth and Woods, they did a lot of work on them. But they didn't understand the equivalent circuit like I did. And so the whole performance of them under transient conditions, they didn't have the slightest idea how they would work under transient conditions. But with equivalent circuit, it was just a voltage back of some capacity. And you'd measure from the terminals what the voltage would be in open circuit, what the impedance was with the voltage short-circuited. And once you knew the equivalent circuit, why, you knew what the performance of the thing was under transient conditions. You could calculate it all. So that's what this paper was. "Transient and Steady-State Performance of Potential Devices." And, again, the other fellows who were working on it are co-authors.
And then capacitor installation to lower the fifth harmonic voltage I told you about. That was on the Potomac Edison system down in Hagerstown — around Hagerstown. And well, that was given at — There's an internal publication. And then there's a paper. And Feaster, this engineer of Potomac Edison. I really thought he had contributed because while he was really just a radio ham, he could go into a station and set up the proper oscillographs and selective circuit to measure and photograph the fifth harmonic voltage. And so he was a co-author on this paper, and I think he deserved it. And that was presented down near Norfolk — Newport News — Virginia. There was a meeting down there.
Then linear couplers for bus protection. I told you about those. Harder, Klemmer, Sonnemann and Wentz are all on it. Well, it was my invention, but they all worked on it, so we wrote the paper jointly. And here's a couple of fairly immaterial ones. They were important engineering problems that the solutions were written up, but there's nothing all that special about them. Now here — "Regulation of AC Generators with Suddenly-Applied Loads." Bob Cheek and I wrote a paper. Cheek did some nice theoretical work on this, and this was really part of working with the Navy on what became my doctoral dissertation later. But this was sort of a preliminary to that. And Bob Cheek did some good work on it. He did really deserve to be on this paper. Hmmm. "Static Voltage Regulator for Rotatrol Exciter." Part of the Middle Atlantic district was Eastern Shore of Maryland. And GE didn't have any office down there. Westinghouse had one salesman down in Salisbury, Maryland. And I had gone down on rare occasions to see him with a salesman. He was a good duck hunter. He'd usually give me ducks; tell me how to cook them and everything. You put a potato in the stomach to take the wild taste out. But one time we had a static voltage regulator on a little machine, and it was in a Chinese laundry. And he took me in there to explain to them [Laughter] what the trouble was. It was the most ridiculous thing you could imagine, trying to explain to the owner of a Chinese laundry why the — You know, if a voltage regulator has contacts that move, you can see it. But there's nothing that moves in a static voltage regulator. Just some circuits that have the same effect, but nothing moves. And to try to explain to somebody that didn't know from beans about electrical engineering — Why that salesman ever took me there, I'll never know. But I laugh still. One of the funny things that happened in my experience: explaining to a Chinese laundryman why the static voltage regulator didn't work. But this paper was on the static voltage regulator for the rotatrol exciter, and apparently I had helped. Valentine and I had worked it out, and we wrote a paper on it. Here's the linear coupler paper number thirty-two.
It's about the field test. The thing I remember about that was these fellows from Metropolitan Edison Company owned these stations down at York and Middletown. Halfway between them was a Pennsylvania Dutch eating place. It was just a big farmhouse, and everything was served family-style. You sat down at a big table, and they kept bringing things and passing them around — all delicious food. And we were young fellows; we could eat any given amount of it. But you know, now I don't eat much. But, boy, then I could really go after it. And so these fellows from Metropolitan Edison, they knew all about this place. We patronized it. And here's my Ph.D. dissertation on "General Solution of the Voltage Regulator Problem by Electric Analog Computer." And the thesis. Then here is 35. As I said, I was always afraid during the war that the two-stage rotatrol might not make it. And I was on a side developing this balanced amplifier using biased saturable core reactors, which I felt I might bring through if need be. But it was never necessary. So after the war I wrote this paper. And as I say, Gordon Brown's students at MIT built it. We had enough data on it they could actually build it. And he was impressed and offered me this job at MIT, which I had to turn down. Then here's a paper on the ANACOM number thirty-six a very fundamental paper.
Now we're getting into lightning. In lightning and relaying both, when I felt I knew the whole subject, I wanted to write a paper that covered it generally. And this was a CIGRE paper; it covered the current knowledge of lightning.... Couple of nondescripts here. At least they had no lasting effect. And this is the relay paper number forty-one that was presented in Mexico City.
Again, by that time I felt I really understood the whole relaying picture in the United States, and I wanted to write it down. And that was it. As I told you, Phil Sporn handed it to Stan Horowitz when he first came there to be a relay engineer. He said, "Here, read this Harder-Marter paper. Phil knew that if he read and understood that paper, he'd have a damned good start on protective relaying because he would have a good grasp of the overall picture. It was all there. Even the figure that gave the principles of the elements — had 26 parts, A to Z. Took two and a half pages for one figure. So I put an X on that one. I see Clayton and I wrote a paper on "Transmission Line Design and Performance Based on Direct Lightning Strokes," but that was basically Monteith's field. He carried Fortescue's work into practical use in designing transmission lines. Here's some new techniques on the Anacom. After the Anacom was built, a lot of additional things we found we could do. And finally we wrote a paper about a lot of them. But that's just a "by-the-way" paper. It mentions books — the Electrical Transmission and Distribution Reference Book — which I was going to show you. I wrote two chapters. One on the steady-state performance of systems, and that gives all the methods of network solution that were used. The elimination of points that I mentioned to you and the different ways of solving networks. And then with another co-author I wrote about relay and circuit-breaker applications for that book. And then later — I think it was in a revision Clayton and I revised the chapter on line design based on direct strokes, transmission line design.
Aspray:
This was a standard reference work?
Harder:
It's been widely used all over the world. I have written in other books and papers, but I'm better-known all over the world for my chapters in this T&D reference book. I run into people in any country — "Oh, you're the one that wrote this chapter in the T&D reference book!"
Nesbitt had just the transmission line constants, and it had to be replaced because there were many new types of conductors and configurations now in use. But Monteith envisioned not only replacing the tables of transmission line constants, which he did (they're in this book), but also by giving the whole fundamentals of power systems in it. So that practicing engineers all over the world and students could see what the current practices were — how all this work was done. And it was widely distributed all over the world and in the colleges and universities all over. And so my appearance in two of the chapters of this book — through those I am much better known to people all over than I am through other books or publications which I have written myself. Or any of my papers. Because this was very widely distributed.
Aspray:
Do you have any sense for what the numbers were?
Harder:
I would just put a figure 10,000. This book probably, 2,000.
Aspray:
The Fundamentals of Energy Production.
Harder:
Yes. Which I should have had the title "The Nine Sources of Energy." But now that there's no longer an energy crunch, why, the sales have practically run out on this. That was published in '82. And there were high sales for a while, but that title is a terrible title. It doesn't tell anything at all. And if I had only put the title — When anybody asks me what it's about, I tell them it's about the nine sources of energy: the five conventional sources are coal, oil, gas, nuclear and hydro. And there are four alternate sources: wind, solar, geothermal and biomass. And I haven't found anything that you can't put in one of those nine categories. I mean, if you talk about tidal energy, that's hydro. If you talk about pump-storage, that's hydro. So every used energy commercially that I can think of goes in one of those nine brackets. And this book has chapters on each of them, as well as a lot of supplementary chapters. The resources of energy, the chemistry of energy, the physics of energy, the transportation and storage of energy. There are chapters on all of those other things, but there are nine chapters on these nine sources of energy. It took years to do all the research and write it and get it published. And in the course of that I visited energy installations all over the United States and a lot in Europe — England and Germany and France. These papers have to do with voltage regulation under suddenly-applied loads, lightning phenomena, capacitors, lightning performance of lines.
And "Principles and Practices of Protective Relaying in the United States," presented in Mexico City. See Number 41 for the actual paper. That's the Mexico City presentation. Then here is a paper in 1952. Now the ANACOM was finished in '48. But this paper "On the Analog Computation of Blades," in 1952, that was one of the first mechanical projects that we had. From the first we had electrical transients, and the transient torques and couplings in machines, mechanical problems. But South Philadelphia had never been able to calculate the higher modes of steam turbine or compressor blades. They had hired a fellow from Johns Hopkins — a student — to come in and make one calculation. And it took several months. And after several months they couldn't be sure if it was right or not, you know. All that calculation. And besides, he died. The student died. So when I went down there to discuss doing it on the Anacom, they told me, "Some of us think this can't be done. And others think that you can do it but it kills you." [Chuckling] The results about the same. So we found that we could do it on the Anacom, and that was one of the first big successes of the Anacom.
Now at the bottom, Number 68, there starts a series of papers, which ended up really with the economic dispatch computer. The first one is "The System Loss Evaluation and Computers," and then "Loss Evaluation 1: Losses Associated with the Sale of Power, In-Phase Method." And then "Current and Power Form Loss Formulas." "Computers: Their Use in Planning a System." But eventually there were good loss formulas developed. Power systems — They know the cost of generation. They know the cost of coal and the efficiency of the units. So from low-load to large-load on any unit in the station, they can tell you what the generating cost is. But when this power goes out over the transmission lines to the load, there is a loss. And if there weren't any other generating stations, it wouldn't be hard to figure what that is. But there's that self-term, and then in addition, station 2 is sending power over the same network. And so there's a mutual term comes in. Station 3 and station 4, up to station 12. If there were 15 stations, there are 15 mutual terms that come in. And the economic dispatch of the system is when all stations have the same delivered cost. So if you had 15 stations, you had 15 simultaneous equations, with 15 terms in each. And they have to be solved. And with an analog computer and a servo on each of these generator cost curves, they will keep changing each station until the delivered cost is what you say you want.
If you want to get a delivered cost of one cent, why, each of these will keep adjusting. And you take into account the component of loss from all the other stations until they are each delivering at a delivered cost of one cent. That's the economic dispatch then. Well, you can do that if you build the computer with servos and with these cost curves set up as resistors or something that generates voltages proportional to these things I'm talking about, why, you can make an economic dispatch computer. And that's what it was. There's a lot more to it than that, but that's the principle of it. And to sell it to the people down in New York, I made a pasteboard model with all the dials on and explained what would show up when you used this and how you would use it and so on. They bought it right away. No question about it. The amount they could save by economic dispatching was so much more than the cost of this economic dispatch computer that there really wasn't any question about it.
So now we come to Volume 4. The oil whip problem. The Anacom was used not only for mechanical problems but some fluid-flow problems. And the oil whip in bearings was one. The research engineers knew how to set up the equations, and we had the equipment; you could implement it, solve it. "Analog for Representing Corona on an Electric Analog Computer." We could introduce any kinds of nonlinearities; if you knew the law of it, we could produce it. And so this would be very difficult to do on a digital computer. But on the analog you can get some little device that you add to the thing, and it takes care of it. See an awful lot here on economic dispatching and loss evaluation. Transient stability studies. That's just another type of study that you can make on a calculator. Oh well, I'll just put a general X here. You'll recognize the economic load dispatching. There are a lot of papers that are connected with that. On the loss formulas and how you use them and so on. Now we're up to Volume 5. And the first paper was the one I presented at Brussels, which told about the six computers that we then had in the Analytical Department, particularly about the Anacom. And at that time a similar machine was being built by ERA in England.
Aspray:
What did ERA stand for?
Harder:
Engineering Research Associates. It was a — Power was nationalized in England, but this laboratory somehow served the system. I don't know whether it was nationalized, but it was called ERA — Engineering Research Associates. It might have been a separate enterprise that served the British network.
Then here's a series of papers that had the general theme of the computing revolution. And I gave it once when Chancellor Litchfield was inaugurated at the University of Pittsburgh. Gave it at different computer meetings, and it appeared in the Association of Computing Machinery Journal. And then IFIP had a tenth anniversary in Amsterdam. And I presented there a paper called "IFIP and the Computing Revolution." Which really tied together all that IFIP had done in the computing field and what its part was in the computing revolution. And Heinz put that in the book — put that paper. He edited the book on the tenth anniversary — Tenth Anniversary Book — for IFIP. And he said, "Well, that paper ties it all together." It wasn't one of the invited papers exactly, but somehow or other I had given the paper there, and he included that one. Told how it had all come and what part IFIP had in the developing countries and how the first committees had to do with languages. How would the different countries talk to each other if they didn't use some common term? So they had to have a subcommittee that dealt with the terminology. And then TC2 was programming. TC3 was education. And then TC4 was the use of computers in medicine, which was very early important. And then TC5 I introduced; it was in technology. Most of the members of IFIP were from schools — from that side. And I was one of the few from industry. So computers were beginning to be used for controlling machine tools. This was industrial use. Started up at MIT. They did a lot of the geometry work — how you describe paths that tools take and so on. And then Westinghouse picked it up. We did a lot of assembly things, program called CAMP. And then a group in Europe took it up, and they had a lot of experts on machine shop technology and planning and programming and all. And they started extending it to that. And so Professor Caraciolo from Italy proposed we ought to have a conference on the use of computers in machine tools. And since I was about the only one that wasn't in academia — I was in industry, you know — they asked me to head up this conference. I got a lot of good help on it. And I visited all over and found out what was going on every place throughout Europe and MIT and different places. And I decided to call it PROLMAT — Programming Languages for Machine Tools. And PROLMAT '69 was held in Rome. We were invited to Rome to have it. And Leslie in England edited the publication. And it was very successful. So there were PROLMAT conferences about every four years for quite a few years. One in Hungary, the next one, and one in Scotland. And then I lost track. But starting with just using computers to control the tool, it got to where it kept track of the whole process and the economics and the cost of doing these things and the scheduling and everything, see. So they kept adding to it. But I was in on the first of it. That was — Well, I'm way ahead of myself here. I see in different papers I would give talks to describe the progress that we were making in using computers for designing equipment and so on. "Computers and Automation" in the Special Jubilee Issue, 75th Anniversary of the AIEE, 1959.
Aspray:
How did you decide which journal to go to? I mean, sometimes it was obvious because you were at a conference of a particular organization. But suppose you weren't and you had freedom of choice.
Harder:
Well, in the AIEE they were technical papers, and they constituted the journals. Now if you wrote a separate article, I guess there was some place it could go in AIEE. But for the ACM they had a journal, and you dealt with the editor.
Aspray:
Right. But you have some publications there that aren't in either ACM or AIEE.
Harder:
If it went in the Electric Journal, there'd be a lot of audience it wouldn't reach. Unfortunately, the paper on the HCB got in Electric Journal. It didn't come out with the timing just right to get the paper to the IEEE, which it should have, because it was probably one of the most fundamental things that I did. But it got in as an article — it was published as an article — in the Westinghouse magazine, which I think was the Electric Journal at that time — later became the Westinghouse Engineer. But the preference would have been to have it as a paper in the IEEE where it's accessible to the whole industry and is on record. Well, the Electric Journal is on record, but nobody would normally use it until after they had exhausted the IEEE material. It isn't likely you would come across it in the Electrical Journal.
Now we're getting pretty well through up to Volume 6. And this is 1959, and I see I was not going to this congress in Paris until I discovered they were going to have sessions on digital versus analog computers. Then I was real interested! And at the last minute I decided to go, and I presented a paper there on digital versus analog computation.
Digital vs. Analog Computation
Aspray:
For a fair portion of time in the Analytical Department you had access to both digital and analog machines.
Harder:
We always had access to analog — even right up to the end. But from 1949 or '50 we had — In '49 we had card machines in the Accounting Department. In 1950, CPC. And so from 1950 on we had digital machines available, too.
Aspray:
When a problem came in, how would you decide which kind of machine to put it on?
Harder:
Well, if it was a little heat-flow diffusion problem, you'd put it on the DC board; you never would program it for a digital computer. The DC board you just practically set it up, you know. If it was a design — like the design of induction motors — it was no question; it had to go on a digital computer. There's a lot of logic involved and a lot of optimizing and a lot of tables and data that were used in the course of the design. That all had to go into the program so there would be no choice whatever. Now one type of thing, the network calculator was very good for getting the solution of the electric network. In a stability study you have these different machines, and their angles are continually changing. And so for a few seconds they're all oscillating relative to each other. But if you obtain first the network solution from the AC calculating board, then you can put the solution in the digital computer and use it for the accelerations of machines and what not and follow them through step by step through the stability studies. So there was a little of that done, but that was very interim because before long you would do the whole thing digitally. But there was a time when the analog machine was still so much better for network solutions that you would get the solution out of the electric networks calculator and then use it in a digital computer for a stability or dynamics problem.
Aspray:
Were there classes of problem that it was about equally good to use the two machines?
Harder:
Yes. As I told you, regulating system problems — First we had the Anacom, and then after the war we had a whole rash of electronic differential analyzers — Reeves, Goodyear, Pace — and these were unquestionably very good for regulating systems. But then Rideout came out with these programs where you could simulate an analog computer on a digital computer. Now you're in an area where you've got to make a decision. But it didn't last long. A few years after that you can get rid of the analog computer because you can simulate it on any digital computer you've got. So why have an additional computer just for a few regulating system problems? So that was a small time interval that there was a question about which way you would do it.
Other things would be somewhat similar. Design for Westinghouse was a big thing and the simulation of systems — big systems. And the design you just couldn't do analog-wise. There was just no way on an analog computer that you could look through a table and pick out a lot of different values. Well, you could later. As on the Anacom that you've seen. Now, Anacom I that I had, you had to set the machine for each thing you wanted to look at — manually. You went and changed the dials. In Anacom II they had digital inputs so you could program and you had machines that did all the setting, and you could have it work all night without your being there. Go through a whole range of constants. Parametric studies of systems. How does it vary? What's the optimum? And you didn't have to sit there all the time and do it. And with Anacom III they have a lot more digital processing of the output data. If you want to get weighted averages or probabilities or something like that, you certainly don't have to look in each case yourself. And you certainly don't have to read it in and out anymore. And you can even take the data and work it up to get the functions of it that you want. So really you have to consider that the Anacom is a special-purpose machine now. It started out as a general-purpose computer. Now it's a special-purpose machine for electric power systems.
But I was talking with Joe Horton over the fence, and I was citing some figures that he just didn't believe — it was impossible. I said, "The value of the electric power systems in this country are thousands of billions of dollars." "Oh!" he said, "thousands of billions?" he said. "That's trillions." He said, "That's getting up to the national debt or something like that." But I had the figures. There's about 600 gigawatts of generation, and they cost a thousand to two thousand dollars a kilowatt. If you say — Well, that's just generation. If you take the whole system, why, $4,000 a kilowatt would be a good figure for the whole system — not just generation. All right, $4,000 a kilowatt is $4 million a megawatt and is $4 billion a gigawatt; each time you're going up a thousand and one. So it's $4 billion a gigawatt, and there's 600 gigawatts. So that's $2400 billions — the value of the electric power systems in this country. And for the world as a whole, it's several times that. So these power systems are so valuable that the cost of maintaining a few Anacoms around to study the changes and the new applications and equipment for them is so trivial that there's just no question about it being worthwhile. People are continually using new devices: DC transmission, more series capacitors, running into problems of one kind and another. Resonance. And they can study them on these Anacom-type machines. And I just can't conceive that those machines are going to be phased out at anytime. Anytime at all. Because the cost of those machines is so trivial compared with this tremendous investment in electric power facilities in all the countries of the world. It's just peanuts; the cost is peanuts compared with what you're dealing with in all these huge systems. So it's a rather impressive machine, and the fact that the Anacom is the first one is interesting. And I think it ought to be written up.
That's as far as I'm going. I'm not going to try to write a tutorial that will explain to a digital man how an analog computer works, believe me.
Further Publications and Awards
Aspray:
I interrupted you from going through your publications.
Harder:
We're up to Volume 6, and there's a lot of repeat here. "Computers and Automation." "The Future of Control Computers" presented several places. Of course that got to be quite a thing. Well, there was always a certain amount of computation in control, but when digital computers became available, then we started putting digital computers in power plants to keep track of all the data — whatever happened — keep a record of it. And to control a lot of it. And they were called control computers. And Westinghouse built a few of them, but we bought most of them. But when you look at the whole control board and all of the meters and instruments and everything, the computer's just a little thing, you know. And that's what we bought, you know. But we built the whole control board for it. And, as I say, we bought the Process Control Division of Hagen, and we had Hagen-Westinghouse, which is still a good operating division of the company, even today. It's over across the river — across the Allegheny River — from here. When we were first getting into that was about the time I hired Paul Lego for the Analytical Department. He's now chairman of the board of Westinghouse. And he got into that particular work, and he told me they did use two of the ones that Westinghouse built for the Seawarren Station. And another station. And then they started using ones that we bought. But he was in that phase of the work in process control, and he got into that. And then he got selected to run our Youngwood Plant on semiconductors. And then the Elmira Plant with tubes. Gradually he went way up through management, became executive vice president, then president, and now he's chairman of the board. So Westinghouse did get into process control and control computers, and that became — [electrical interference sound] We added a central computer to the Analytical Department because our engineers did a lot of the programming that was needed. Now there was an organization called PICA, which meant Power Industry Computer Applications. And so a lot of these uses of computers in power systems were discussed in the PICA conferences. And I had Professor Humphrey Davies over here to give the principal talk at some of those. And I visited him in England. I did a lot with the PICA conferences over the years.
And then here I see the Lamme Award. One of those medals is the Lamme Award. And one fellow gave the history of the Lamme Medal, how it got started. And Monteith presented the award; told the career of the medallist, which was me; and I gave the response. And I satisfied a long-awaited opportunity tell what a wonderful [manager he was].
So that's the Lamme Medal. It's given for "Meritori Achieve in the Development of Electrical Machine." I had a good friend, John Calvert. Well, let me tell you about Calvert because when I graduated from the student course and got into power engineering, we used big double desks, and Calvert sat across the desk from me. Well, Calvert was a little older. He'd been in World War I. When he got out of World War I, a good buddy of his wanted him to go in with him on a little magazine where they would take the most important works out of other magazines and digest them for this one. But Calvert's father had been a professor. He wanted to be a professor. That was his goal in life. So he reneged and the little magazine became The Reader's Digest. The fellow that started The Reader's Digest was his pal in World War I. Well, Calvert went out to Iowa, and then he went to Northwestern. And while he was at Northwestern he had contracts to study Army Ordnance. And we supplied an Anacom to them for about $300,000. It'd be about $3 million today.
But he knew I had built the Anacom here, and he wanted a similar one at Northwestern to work on the problems of armament. That he was helping the Army on. And then he came back to Pittsburgh and headed up the U. Pgh Electrical Engineering. And when he came back he had certain goals, certain things he was going to do. And one thing, he was going to get that Lamme Medal for Lee Kilgore, and he was going to get one for me. [Laughter] And so he got the information someplace — probably from my secretary — and Lee Kilgore got the Lamme Medal, and then I got it. But I had a chance to thoroughly air out what I thought of the Westinghouse management, particularly Monteith. All he'd done in education in every way, he'd been so supportive. So that's the story of Calvert. He was a good friend, but he also was the one that nominated me for the Lamme Medal. And he had the Anacom out at Northwestern, so he knew quite a lot about me, and he knew me from way back because when I was two years with Westinghouse, the second year I'd sat across the desk from him in Power Engineering. He was a pretty good friend.
I see this presentation of the Harry E. Goode Memorial Award. That was the one to Howard Aiken. It's not important; it's just mentioned. I'll put an X on it. 150 is marked IFIP Congress '65. And of course I was chairman of the Board of Governors of AFIPS at that time. I think the name "president" was adopted the next year, but at that time you were called "chairman of the board." Not much of importance there. Now this article, "The Expanding World of Computers," got into the Communications of the ACM. Well, I was a member of the ACM, and I could submit it all right. But I think it was easier to go that route than to try to get into any IEEE general publication. Other than technical papers, I would have always presented to IEEE. But this was an article that was of general interest. And the acceptance of the AFIPS Distinguished Service Award at Atlantic City; that's that award — 1971. And, well, I had been president in 1965, so that wasn't too long afterwards. But I had not only been president, but I had been the United States representative to IFIP for several years, and so it's sort of a joint reason for that. And, as I say, this article on "IFIP and The Expanding World of Computers" — Heinz published it in the IFIP Tenth Anniversary volume called The Skyline of Information Processing. Because he said, "That ties the whole thing together." I told how IFIP handled all these different things. And the others had treated them one by one, you know. And IFIP finances — I was the treasurer. Then this book that you looked at. There are three chapters I was involved in, and you've already seen those. Master's thesis. The Ph.D. dissertation.
"Specific Output of Windmills." That was a very interesting thing. It had always been assumed that the way the wind varies throughout the world — on the seacoast and mountaintops and so on — that for any windmill you had to know the characteristics of it and know the annual distribution of wind. And then you had to go through more or less hour by hour and see how much output you got. And I discovered that you didn't have to do that at all. All you had to know was the mean velocity of the wind and the speed at which the windmill would come up to full power. And from that ratio there was one curve that went through all the points that were available in all the literature. Twenty-three stations that they tested in England — all different places. A station that the Department of Energy, they had some windmills they'd worked out for the United States. I put them all on. They all fitted on one curve. It was just a plain discovery. And this meant that all these other things were fairly unimportant. So that wasn't an invention, that was simply a discovery. And I sent it to the Department of Energy, and they said that they would certainly use it because you didn't have to calculate it all. All you had to know was what was the mean annual velocity of the wind and at what wind speed that windmill would come up to full power. If you knew those two things and take that ratio, you could look on this curve and you'd come within 5 or 10 percent.
Well, that's a lot closer than it is from year to year anyway. It varies a lot from year to year, but this was plenty close for anything where you needed to know the specific output of a windmill. And I had an English publication — by ERA, incidentally — in which they had investigated 23 locations in England and had calculated what the specific output of windmills would be at these locations. And I took the mean annual velocity of those locations and the speeds. And I just put all those points on the curve, and they all fell right on this smooth curve. It was a very interesting discovery. At Plum Brook. in Ohio, they had a windmill. And the fellow that was running it was going to come to Pittsburgh and give a lecture at the University of Pittsburgh. And I thought, Well, I'd better do a little work on my windmill chapter before he comes. And I made this discovery before he arrived. Within two weeks I had — Of course I had these publications available, but why nobody in all the history of windmills had never discovered that before. They all assumed that you had to calculate it for each different kind of wind — on the seacoast or the mountain or trade winds or whatever it was that would be different. So when he came, I didn't get to discuss it with him then. I went out there later and visited their installation at Plum Brook and discussed it. And the engineering was done in Cleveland by some group. I'll put an X there. That was an interesting part of my work on energy. I certainly didn't expect to make any discoveries about windmills, and here I did.
"Sailing a Motor Cruiser." I have another model of that boat upstairs with the sail on it. I think I mentioned that to you. I won't take anymore time on that. Now we're getting into — There were a lot of things that I'd written over the years that I didn't want to just lose, so I put them in here. They're not technical things. For example, in 1981, ten years ago, I planted eight dozen tulips. And I wrote a story about the 96 tulips for Esther for a Christmas present. And that's in here — that story. So that's it. That's the papers. I have another whole volume — will be volume 8 — but it isn't organized yet. And so I may have one more bound before — if I live long enough. The eighth one will be a lot more personal and hardly any — Well, there's some technical things in it. What was it I discovered the other day that — ? Oh, I wrote the history of the Anacom, which I think ought to be published. And it isn't in any of the first seven. It's the one I gave you. And that will be in the eighth. So there'll be a few technical things in the eighth, but largely [electrical interference sound] a write-up, a collection of things.
In this folder I'm collecting doubles, so one set goes to one of my sons. And the other set I have here. I don't know what'll happen to it when I die. But when I get maybe three inches of stuff, why, then I'll organize it and get it bound. Probably Volume — Well, this is 6. Well, the index went through 7. Volume 8 will certainly finish it as far as I'm concerned. But my secretary started doing this when I had the Analytical Department. She started collecting my papers and got the first few volumes bound. And then I [electrical interference sound] continued it after she was no longer my secretary, and it's very useful rather than have to go look things up. If I know what year it is, I can pull it right out and find it.
Most Important Contributions
Aspray:
In just a few sentences, could you go back over your career and tell me those half dozen things that you think are your greatest accomplishments?
Harder:
Well, not being chronological, certainly the Anacom and the Analytical Department are my top achievements with Westinghouse. There's no question about that. And that's what I'm known for. It was called the Advanced Systems Engineering and Analytical Department, but to everybody it was the old Analytical Department. That was it. And I was the head of it. And all these people that know me — the ones you met today, some of them — they knew me from the Anacom and the Analytical Department.
Aspray:
If you were to name a few of your patents which ones would you point to as most significant?
Harder:
Well, the patent on using the single-phase directional elements for controlling carrier, which resulted in many millions of dollars of business to the company; that's certainly one. The HCB relay is one of the best patents. The economic dispatch computer should have been, but it was mishandled. It only resulted in one computer, and it was a shame. The linear coupler was a unique patent, a unique solution — a real clean-cut, unique solution to something; it was very valuable. The rest I would — Each played its own small part in a big picture, but those particular patents were worth quite a lot. The HCB relay and the carrier current relay system, which is still used — both of those are still used — today. This is from 1940 to 1990 — fifty years later. There hasn't been any better way found for doing those jobs.
Aspray:
How important was your work to the success of Westinghouse during your active career?
Harder:
Now that is hard to say because as you'll read here, a railroad electrification — a successful railroad electrification — or a successful computer like the Anacom is never the brain work of one person. And it depends on having associates who can carry the ball so well that their combined efforts add up to success. The combined work of me and my associates was very important. The Pennsylvania Railroad electrification was. What I did contributed, but there were a lot of very capable associates. I was lucky to be in Westinghouse where they existed. The Anacom I didn't even have the original idea. It was McCann's idea. But I had extremely capable people all the way through to assist me with it. And I'll tell you a little story about that. When I had been with the company for 35 years — it would be 1961 — the members of my department, along with the members of the Electric Utilities Department which I had left upstairs with which we had close relationships, and the large rotating machinery people in East Pittsburgh that we were sharing the computer with actually at that time, and some of our neighbors in the "L" Building, all got together and had a testimonial dinner for me. A surprise testimonial dinner. I didn't know about it until I got out there. So far as I know, I am the only manager that has ever had anything like that. So I've had all these awards, you see, but now that spontaneous upwelling of goodwill and friendship on the part of all the people that I worked with is more important to me than all of these.
Aspray:
Let me put aside the question of individual credit versus team effort for now. Tell me, more specifically what the Pennsylvania Railroad and then the Analytical Department's efforts just what they did for the company.
Harder:
Well, of course the Pennsylvania Railroad, the fact that General Atterbury had faith in the economy of the country as president of the Pennsylvania Railroad and went right ahead with it all during the Depression, pretty much saved East Pittsburgh. Right in the middle of the Depression we had a $10 million order for locomotives in East Pittsburgh. And it almost carried the plant during that time.
Aspray:
How about the company as a whole?
Harder:
Well, there were hard times all through the company. A great many people laid off and all. But not many parts benefited as greatly from the Pennsylvania Railroad as East Pittsburgh did because we built the locomotives there, as well as the switch-gears. Well, the transformers came from Sharon so they benefited-relay from Newark.
Aspray:
In all those different components of the Pennsylvania Railroad work, what portion of the company's business would they represent during the Depression?
Harder:
Certainly over 10 percent. I think $100 million might have covered the company's business in that year that we got a $10 million order for locomotives. And so it would be in the 10 or 20 percent bracket.
Well, of course I would have been laid off it hadn't been for that because Pete West came from Switch Gear and he said, "Ed, don't worry. As long as I have a job with the company, you have." And so it was clear that he was going to take me from General Engineering into his department to handle — continue to handle — the Pennsylvania Railroad technical work, which I was handling at that time. And so I guess I owe my stay with the company to the fact that the Pennsylvania Railroad went right ahead with that electrification.
Aspray:
What difference did the ANACOM make to the life of the company, to the health of the company over time?
Harder:
Oh, very little. I mean, they would have muddled through with all those things. We were building the Tullahoma Wind Tunnel blades. Well, it was a big help to be able to calculate the higher natural frequencies, but they had a way of building the blades. They'd build small ones and test them and see what the frequencies — And then they'd build the larger ones. Extremely expensive, costly way of doing it. So this saved a lot of money and added to the profit — profitability — of the operation. But I can't think of anything that we did that was all that vital that the company rose or fell on it. Or that you'd have trouble pointing to any difference in the profitability of the company from that. The 1936 carrier relay invention certainly increased the profit of two departments, Relay and Switchboard, very materially for several years.
Aspray:
Would you say that the work you were involved with was more significant to the company or to the engineering profession, engineering knowledge, as a whole.
Harder:
It was more important to the company. The way Paul Lego puts it, he cited: "His technical achievements as instrumental in increasing the corporation's stature in engineering and innovation." In other words, I was one of many associates that made Westinghouse well known as a good engineering company. I did my part; so did many other people. And then he said, "During his tenure with the company, Dr. Harder was a strong force in the development of young engineers as designers of electrical apparatus." Well, I was thinking about that, and I hadn't realized it, but I guess that's right. I worked with a lot of engineers, and I certainly helped them. Like teaching — I did a lot of teaching on circuit theory and design of nonlinear circuits and things. So I guess I did help them. And he was saying: "He was honored for his dedication, his engineering achievements (which we've discussed thoroughly), his educational accomplishments (I've told you a little about that; I taught Pitt-Westinghouse courses quite a lot and apprentice courses), and his guidance and inspiration to others."
Aspray:
If you were to point to some things that had industry-wide or profession-wide value, what specific things that you contributed would you point to?
Harder:
Well, I think the professional society activities would fit best in that category: the contributions to IEEE and to IFIP and AFIPS and all were certainly industry-wide things. It's a way that we have of sort of unconsciously sharing large developments. In the IEEE all of the current contributions end up in papers and are put in the board for everybody to know about. And they are the inspiration and starting point for the next group of papers. So in a sense we are carrying out this development of electric power systems, or whatever you're working on, jointly. All the engineers are working jointly and contributing their efforts to this pool. And we're pooling it and teaching each other and advancing much faster than if we ever tried to keep that all secret to ourselves. So it's this cooperative development through joint sharing of things in the industry-wide IEEE and similar organizations that is responsible for the huge development of electrical engineering and computers and electronics and all that in this country. If we didn't have that, it would never have developed as fast as it has. So I think that my part in the IEEE and IFIP and AFIPS and all is contributing to this overall advance more than my specific contributions in just what I did in Westinghouse.
Aspray:
What importance did the work of the Analytical Department have in the furthering of general knowledge of analog computation or, more specifically, its application to power?
Harder:
Well, it led the way. Now Bob Evans and Monteith had built this electrical transients analyzer. But they couldn't set up whole systems. So with the Anacom we had enough elements that we could set up whole power systems and study these things. So this made it possible. And by the time that I was leaving that work and Westinghouse got the contract for the design of the 500 KV system at VEPCO and it became clear that switching transients were important parameters and they designed circuit breakers with closing resistors, why this method — they could never have done that at all without the Anacom and these methods of studying systems.
Aspray:
What difference did it make to people at other institutions or in other countries? Was the work in your department watched closely by others?
Harder:
Undoubtedly. The Anacom was the first, 1948. But by 1955, ERA was building one in England. And I was just looking at a letter from this Anacom group, and they were pretty sure that CEZI has one in Italy. That's the Edison group in Italy. And there's one out — McGraw-Edison has one in Pennsylvania not far from here. And GE does have a machine like the Anacom now. They didn't have for years. If they wanted to do that kind of work, they'd collect the equipment and hook it up and get some experiments. But they didn't build it like a computer like we did. We elected — I suppose that was one of the things that was responsible for the overrun in costs but we elected to make a real computer out of it right from the start. That was a conscious decision. Because that had not been determined when McCann left. And I suppose being young and immature, it never occurred to me that this was going to affect the cost. I found out later that it did.
Assessment of Career
Aspray:
Are there things that you either view as a failure in your career or things that you would have liked to have seen go another course?
Harder:
Well, this economic dispatch computer should have been commercialized. No question in my mind that that was a bad mistake. Part of the responsibility is mine. I was young and immature, and I should have gone to bat for it. But it never occurred to me that this was an excellent product that should be used. I just did my job getting that one designed, built and installed and then forgot about it. I had other things to do — I was busy. But looking at it — back — that was a bad mistake. All of the things that we did during the Depression on metering compensators, of which there are a number of patents and all, it turned out that they were transitory. That the power companies didn't want to do those. They only did it because they didn't have any money.
Aspray:
But they met a need at a special time.
Harder:
They met a need at the time. They were able to do things with less money. And later when they could get the money, it was just better for them to do it simpler.
Aspray:
How would you like to be remembered?
Harder:
Well, as I told you, the spontaneous upwelling of friendship and respect on the part of all my associates is the way I'd like to be remembered. Not by these patents. The first 20 years were in electric utility engineering. And then I got over into computing. And my career from there on was closely associated with computing. Then why did I go back into power and write a book on energy, you know? Energy was my — that was my first love. Well, the beginnings of that I was always interested in energy in cycles: the way the cycle that you use in an automotive engine, the cycle you use in a steam turbine, the cycles that you transform energy from one form to another. These always interested me. And then I was interested in: What is the heat balance of the United States? Where does all the energy come from? What's it all used for? And I was — before I retired even — I was making tables of this, and I was comparing with the vice president of Engineering. He was a little interested, too. We both had this interest. But from energy in cycles, which was the starting point, and then from wanting to have a broader idea of the use of science in the world, is the reason I retired in 1965 — although I did go back half time as senior consultant. I learned that some things you could get a pretty good idea of in a short time. Like the population of the world. Who are all the people? And the food of the world — particularly from an energy standpoint. What was the food of the world? What do people eat? And what were all of its food values? They're in this book. Astronomy. What does the universe consist of?
If you're willing to spend a few months or several months, you can get a good idea of it, but modern physics you can't. There's no shortcut to learning modern physics. And so I reached this conclusion, and I just got a physics book by Sproull at Cornell on modern physics, and I read it clear through and I worked every problem, and I learned modern physics. And it's a good thing I did because later on I was dealing with nuclear energy and trying to understand the developments at Siemens and at Westinghouse. And if I didn't know modern physics — When I graduated from college, the neutron hadn't been discovered yet. That was in '32, I think — several years after I graduated in '26 anyway — that that was first discovered. And so all modern physics was after my college training. And then the more I studied them, the more I got into — Oh, first I thought that all alternate energies I could cover in one chapter: wind, solar, geothermal. Ends up there's four chapters. There was more to each one than I thought. I think that's all I want to say was that's how did I get from in 1965 wanting to get a broader idea of science and even when I went back half time as a senior consultant, it allowed half time to study. And I did study at the Sorbonne in Paris. I studied in Munich at the library there and learned a lot of these things that I wanted to know. And at the libraries in Pittsburgh. And modern physics wherever I was. I just finally waded through Sproull's complete book until I understood the whole thing and I could — I told him a lot of errors that he — He said, Well, he didn't know whether he had another revision in him or not, but he'd put it in the file. And if he got to another revision, why, he'd use it.
Westinghouse Publication Policies
Aspray:
When you were preparing publications while you were an employee of Westinghouse, were publications screened by the company in any way or was there a formal procedure for publishing?
Harder:
Only the presentations. Before you made a presentation to the IEEE, there was a practice session. And you were criticized and screened. And that was terrible! I just couldn't stand an adversary criticizing me. But I got up in front of the IEEE, they were friends, anxious to hear what I had to say. And it was just like somebody sitting down at my desk and talking to him. But to get up and give a speech before these adversaries and have them ruthlessly criticize. What I'd say, it wasn't until I was able to throw that out of my mind and just treat that as something you had to go through — that was the company rules, you had to go through it — and then I talked the way I wanted to talk. Not from some outline — that you made the outline and filled in pieces and then all this monkey business. I talked the way I would talk. And that's the way I would've talked to somebody if they'd come to my desk and sat down and wanted to know about something. And that worked. And so that worked for me in making presentations. That's the way I talked. So that this business of the company monitoring how you were going to deliver it and all that — for me that was just a horrible nuisance.
Aspray:
Did the company ever restrict publication? Was there proprietary information, for example?
Harder:
No, I don't ever recall getting into that. I'm sure that there were things that you weren't expected to — As an employee you were supposed to know what you couldn't say — whether you were out talking to a customer or what not. There might be weaknesses. You weren't about to point those out.
Aspray:
But you were supposed to be self-policing on those?
Harder:
Yes.
Aspray:
Okay. Well, thank you very much.
Harder:
Not at all.
- People and organizations
- Corporations
- Engineers
- Inventors
- Research and development labs
- Profession
- Business
- International collaboration
- Power generation economics
- Defense industry
- Power industry
- Innovation management
- Research and development management
- Engineering and society
- Military applications
- World War II
- International affairs & development
- Law & government
- Patents
- Professional societies
- IEEE
- Governance
- Committees
- History & heritage
- Publications
- Computing and electronics
- Analog circuits
- Analog computers
- Calculators
- Engineering fundamentals
- Mathematics
- Systems engineering and theory
- Energy
- Current control
- Phase control
- High-voltage techniques
- Power system simulation
- Power generation
- Hydroelectric power generation
- Nuclear power generation
- Power distribution
- Power generation planning
- Power grids
- Power supplies
- Power system dynamics
- Power system economics
- Power system faults
- Power system harmonics
- Power system management
- Power system measurements
- Power system planning
- Power system reliability & stability
- News
