John Parsons has been an SME Fellow since 1986, and an Honorary Member since 1998. In 1975, SME honored Parsons with an Engineering Citation as the person whose “conceptualization of numerical control marked the beginning of the second industrial revolution.” In 1998, the retired founder and president of the John T. Parsons Co., Traverse City, MI, was honored for his 70-year history in manufacturing and his contributions to the automotive and aerospace industries. Here is the story of his early life in manufacturing and his account of how NC developed.
By Russ Olexa, Senior Editor
In the late 1940s, John Parsons started a revolution that engulfed industry and continues to be of major importance in just about every person’s life.
His invention cut manufacturing costs, added to society’s comfort, strengthened the national defense, and today supports a standard of living envied by the nonindustrial world. The Society of Manufacturing Engineers calls him the Father of the Second Industrial Revolution, and his work has truly made him that.
As Parsons explains it, necessity was the mother of invention when he tried to solve a problem with the help of an employee, Chief Engineer Frank Stulen, and eventually patented numerical machine tool control (NC), the forerunner of today’s digital computer-control systems. But before he patented the system, there was controversy and intrigue.
Who is John Parsons, this self-taught engineer, and what were the events that led to the development of NC? In an exclusive interview with Manufacturing Engineering, he discussed his early life, and the events that led to this revolutionary invention.
Russ Olexa: How did your life in manufacturing begin?
John Parsons: My dad came to Detroit in 1911 and started a stamping company, called Parsons Manufacturing Co., in a small factory on Detroit’s 21st street. In the 1930s he expanded several times, and had a larger facility located in the back-end of the Ford Highland Park plant with about 34,000 ft2. His main business was stamping automobile door hinges, door hardware, and window lifts. Within ten years he built that business to 565 employees.
I started working in my dad’s factory in 1927 at age 14 as an assembler, and then I worked as an apprentice in the toolroom. After awhile I was doing die repair and try-outs. In 1931, the company’s gross sales were $28,000. In 1940 we had grown to $700,000, which was pretty good growth.
RO: How fast did you progress in the company?
Parsons: By the time I was 20, I was general manager of our automotive division. I was handling sales, which kept increasing. My wife, Betty, and I were married in 1940 and then WW2 came along. I had some physical problems that were not properly diagnosed, and I wasn’t going to make a good soldier. I was determined to get into war work, and that happened when I saw an ad in the paper looking for people to bid on a defense project to make landmines, and I submitted a bid. When they opened them, I was shocked. I was low bidder on a high-explosive mine by 11%, and on a practice mine by 22%. I told my father I must have made a terrible mistake because we got the job. He said, “Don’t worry about it, we’ll look it over.” Well, I hadn’t made a mistake. I just knew what I was doing. By manufacturing 240,000 landmines a month in our Detroit plant in 1941, we became the largest producer in the country.
Later, we bid on entire bomb casings, which included 100, 250, 500, 1000, and 2000-lb ones. We made a sample bomb and got the contract for 10,000 a month. We had to build a whole new plant, and that’s what brought me to Traverse City, MI in 1942. We had just gotten into production, and then the government started adjusting production schedules.
The first NC machine tool, which was demonstrated at the Massachusetts Institute of Technology in 1952.
After negotiations, I ended up building 15,000 a month.
Once we were into production, the Army decided to cancel the program. So I had to go to Washington and find out what was going on, because I had a plant, and nothing was in it except production of the bomb casings. I found out they were going to need fragmentation bombs. I designed and engineered a complete fragmentation bomb, put it on a little hand truck, and dragged it around the War Dept. to get a contract and got one.
RO: How did you become involved in aerospace work?
Parsons: In 1942 Bill Stout, the chief engineer for the Ford Tri-Motor airplane, and I got acquainted. He told me that no airplane with a take-off or landing speed of less than 50 mph would be successful. Then the Detroit Free Press reported that Stout had seen this new invention, the helicopter, the airplane of the future. I called Stout and reminded him what he had said about a 50-mph landing speed. He said, “John, this is something different. Look into it.” So I put in a call that day to Igor Sikorsky, of Sikorsky Helicopters. I ended up talking to the chief test pilot, Les Morris. He said if I was ever in the area to see them. Well I made sure I’d be in the area, Bridgeport, CT, in three or four days.
Sikorsky was setting up a new plant in Bridgeport to build helicopters. I went there, and I was introduced to the purchasing agent. I told him that I wanted to make something for the helicopter—I didn’t know what—and I had no aircraft experience. The only aircraft-related manufacturing experience our company had was when a barnstorming pilot came into Chicago in 1909, and needed a new propeller for his plane. At that time, my father had a shop making wooden automobile bodies in Chicago. Dad carved out a propeller for that man. Well, the purchasing agent said, “We’re kind of busy right now, give me a call in a couple of weeks.”
RO: So it sounded like he didn’t need your help?
Parsons: Two weeks went by, and I hadn’t heard anything. I was in Washington and I called the agent again. He said, “This is amazing, I just signed a letter asking you to stop in when you can.” I said, “I’ll be over tomorrow.” Then he replied, “We’ve made a decision, we’re going to sub-contract the rotor blades, and I told them about your propeller experience.” They had an engineer interview me and approve me as a source for the rotor blades. They introduced me to Boris Labensky. Well, Labensky and I talked for 10 minutes, and I knew I was going to be in the rotor-blade business. The rotor blade was built with a step-tapered tubular steel spar with plywood ribs and fabric skin. I didn’t know anything about it, I knew stamping. They had built about 40 blades in their experimental shop. I said, “How much are these blades costing you? We’d make them for the same price.” At that time they were $750.
When I got the contract in 1944, I had to get another plant in Traverse City. At that time, the other plants in Traverse City and Detroit were being used for auto-parts production. I found an old furniture factory with three 12 by 12' bays and 7' ceilings. We had to make a blade with a 24" chord and 22' long. It was Sikorsy’s R-5 blade; this number was also the helicopter’s model.
Giddings and Lewis (Fond du Lac, WI) developed the first commercial machine tool around 1955. This five-axis machine used a magnetic-tape drive and had a 6–18' movable table and two headstocks, each with motion along two axes, giving five axes of motion in all.
Sikorsky was making blades by hand. I made assembly fixtures. Blades were complicated because a balance weight was installed on the front end of the blade to help with the tilt up for the pitch-change angle. My assembly techniques worked out great. Then Sikorsky wanted to increase the production to 410 a month, so I made 10 assembly fixtures for this quantity.
One of the first 18 blades we made failed on December 1, 1944, and the pilot was killed. It turned out that it was a design flaw in the spar that caused the failure, and not the blade itself. This accident put Sikorsky’s business on hold. The chief metallurgist at Timken Roller Bearing had pointed out the flaw in the metallurgical design of the spar. The steel tube broke so, of course, the blade went.
RO: What did they expect you to do?
Parsons: I heard about this when I was coming back from Bridgeport, CT on a Saturday, and stopped in the Detroit plant. In the few minutes I was there, I got a call from Labensky. He said, “John, we’ve had an accident. They’ve got the inboard part of the blade at Wright Field, Dayton, OH [now Wright Patterson Air Force Base]. I’d like you to go down there and talk with them.” Labensky said they were having a meeting at Wright Field the following Tuesday, and asked if I could be there. Well, of course I could, but I thought about it, and made up my mind to have a solution by that time. I called the head of the Chrysler Cycle Weld Division, Bob Morrison, and told him that I was coming in with some samples on Monday morning. I wanted them to adhesive-bond the metal collars on the spar to the blade, which was never done before on primary aircraft structures. They had been spot welded before, which contributed to the failure. I wanted him to work on these samples, and I told him that I would be there at 8 am. Next, I called Auto City Metal Spinning Company. I had decided on two designs of collars to bond the blades to the spar and eliminate the spot welding that was being done. What I designed was just a cup. I made two different designs, but I had to have the cup to correspond to the diameter of the spar at each step. On a Saturday I called Auto City Metal Spinning, and persuaded them to call in some men to make the cups. Then I had 16 diemakers come into our Detroit plant on Sunday to finish these cups into the shapes we needed. I called our Traverse City plant, and gave them a bill of material of things that we would need, because I was determined to go into this Tuesday meeting with samples of my designs.
The first all-composite airplane built for a Mississippi company by Parsons’ company. John Parsons stands to the far right.
They brought all those samples down Saturday night. We had two-foot sections of rotor blades completely made up, except for the fabric covering. I brought all these samples to Morrison and he said, “You said you were going to have a few samples, and look at all this stuff. We’re not going to get this done today.” Well it took me a half hour to convince him he was. So we left there at 7 pm on Monday and drove to Ohio. The next day, we got into the test lab at 8 am, and tested the parts. We walked into the meeting with data and samples. We were done by 4 pm and Michael Gluhareff, the chief engineer, came to me at the end of the meeting and said, “John, you’ll never know what this means to Sikorsky.” We had a demonstration run on the 3rd of January 1945; Sikorsky was back in business as of that day.
RO: What was the next step in designing the blades?
Parsons: After manufacturing them, I thought there had to be a better way to design a rotor blade than doing it like Sikorsky. Instead of all these pieces, I thought that I could stamp out a rotor blade. I went down to Wright Field to see Frank Stulen, who was the head of the Rotary Ring Branch at the Propeller lab, part of the Air Corps at the time.
He knew about the quick fix on the Sikorsky blades, but I called for this meeting on my metal blade idea, and 10 minutes after my presentation on the metal blade, I sensed that Stulen had realized that I didn’t know what I was talking about. I made up my mind that if I ever decided to build an engineering department for an aircraft division, he would be my chief engineer. So I approached him and asked him if he was interested.
RO: How did he react to the invitation?
Parsons: We got together and he started working for me on April 1, 1946. We interviewed three other engineers and hired them, so that started the engineering department. Then Stulen got the idea from his brother of using an IBM punch-card reader to determine the stress levels on rotor blades, because there was so much calculating time required. His brother, Foster Stulen, was chief engineer of Curtis Wright Propeller, and he had the idea of using punch-card accounting machines to speed up all the design calculations. Frank worked with the punch cards, and I asked him to figure out the program and go to the Grand Rapids, MI IBM service bureau and see how it works. Frank and Jim Gean, another engineer, had 154 steps in their program, and Frank came back and told me there were only two mistakes in it. I said, this is a go. What equipment do we need? We had to rent the IBM machines, and we needed seven for the work. When I saw what Stulen was doing with these machines, I thought they would solve the trouble we were having with templates for the helicopter rotor blades. Sikorsky had furnished the templates, which was standard practice during World War II.
RO: What were the manufacturing problems involved in making an airfoil template?
One of five P-38 war planes bought by Parsons (next to propeller).
Parsons: To define an airfoil template, they gave us 17 points between the radii on the upper and lower surfaces. The coordinate points were different for each of the two surfaces. Then you had to take a French curve and connect those points. It wasn’t accurate, and you didn’t know the accuracy of the French curve between the coordinates. So I asked Stulen if he could use his college education, something I don’t have, and give me 200 points along the radius of each surface, which he had no problem doing. He made up a chart describing X axis and Y-axis coordinates for a milling machine. Then, using a Bridgeport mill, we put one man on the left-to-right axis, and one man on the in-and-out axis. We didn’t have to worry about the Z axis because we were using a 0.050” thick steel template. That’s the way we made templates in the late ‘40s. That’s what prompted the work to do it automatically.
RO: How did the Air Force get involved with your process?
Parsons: My salesman, Robert Snyder, was visiting Bill Wilcox at Wright Field, and Wilcox said they were having a terrible time with a new jet plane’s wing construction. The Air Force had given a contract to Republic Aviation for a fighter plane, and to Lockheed for a medium-range bomber with a 220' wingspan. The designs were so heavy that the planes couldn’t fly. They were wondering what to do, and they called me to see if I could figure out a way to build them with a machine. So I talked with Wilcox and Lockheed, and went to California. The first day there I spoke to 50 different people and only one, a fellow by the name of George Papen, showed any interest in a machine. That was on a Monday morning. I called him the next day and said, what in the world is wrong here? He said, “John, you’ve thrown them for a loop. Give them a few days, come back in on Thursday.” Well I still hadn’t made any converts except for Papen. Now just picture the situation for a minute. Lockheed had contracted to design a machine to make these wings. This machine had five axes of cutter movement, and each of these was tracer controlled using a template. Nobody was using my method of making templates, so just imagine what chance they were going to have of making an accurate airfoil shape with inaccurate templates.
Because of this work, and an article on it in Business Week magazine, a general wrote me, saying they would be interested in financing the development of a machine to make templates. In December 1949, Stulen and I prepared a dog-and-pony show for the government’s visit using a Swiss boring mill located at Snyder Machine & Tool Corp. in Detroit. It turned out there would be about 10 people from the Air Force and one from the Navy to see the process. Stulen calculated the angles for a template. The group saw what the machine could do with the coordinate points and the leadscrew. I received a $200,000 contract from the Air Force Air Material Command to build an automated machine, which of course, was grabbing a figure out of the air.
The director system was built by MIT’s Servo Lab, along with the specially designed paper-tape key-punching desk. Note that the input to the director is paper tape; output signals are recorded on the magnetic tape unit shown at the right. The system was advertised by Giddings and Lewis as their Numericord.
RO: Were you going to build the numerically controlled machine yourself?
Parsons: At this time, I had already started working with Snyder Machine, which manufactured transfer-style machines, to build the equipment without the Air Force contract, and I had incurred about $150,000 in expenses with Snyder Tool.
Snyder had designed a machine, and then Stulen and I realized that we needed servomechanisms to get accurate positions. Massachusetts Institute of Technology (MIT) in Cambridge, MA had a servo lab, and I went to see them. I gave them a subcontract to design the servos for the machine. The contract originally called for research from July 1, 1949 to June 30, 1950. It was extended to February 1951. At this time we called my machine the Card-a-matic Milling Machine. When I got the contract from the Air Force, I immediately hired a fellow by the name of Robert H. Marsh who was with a New England machine-tool builder. He was the one who talked me into going to MIT because of their servomechanism lab. I hired Marsh as the liaison engineer with IBM, because I had signed a contract the previous December in 1948 with them to furnish my company with a data-input device for their punch-card machine.
RO: Did MIT stay on track with the servomechanism portion of the machine?
Parsons: The problem was that MIT overshot their budget with me by about $50,000. I finally had to ask the government for more money for the servomechanism. MIT gave me an amount that I used for the bid then I added on my portion of the work. So I put in my bid for a price increase, and MIT underbid me.
MIT told me they had one overhead rate for private industry, and another, lower one, for the government. But I never dreamed that anybody as reputable as MIT would deliberately go ahead and take over my project. MIT knew the costs were going up, and they were afraid the government might back out on the whole deal. They were even looking for a machine to experiment with. MIT was aware of a lot of government surplus machine tools used during World War II, and they went looking for one, which ended up being a Cincinnati Hydro-Tel vertical milling machine with a 24×60" bed size.
At this point, MIT negotiated a new contract with the Air Force that essentially removed the Parsons Company from further development of the NC system. Originally, Parson wanted to build two Card-a-matic Milling Machines, numbers one and two. One was an experimental working model, and the second model would be the final machine. The experimental machine would have an 18×40" table with the final machine having a 28×72” table. The machines would use an IBM punch card reader that fed cards into a standard IBM calculating machine. Next, the calculator would deliver pulses to a servomotor to move the ballscrew. Later the Parsons team found that the card reader would be far too slow to achieve the goal feed rates of 15 ipm, and Marsh suggested a paper-punch tape or a magnetic-tape reader. A paper punch-tape reader that used a keypunching machine to put holes in the tape was incorporated into MIT’s final design on the Cincinnati milling machine. A digital processor using vacuum tubes took the place of the IBM punch-card calculator that Stulen originally envisioned to control the movements of the machine’s axes.
How did the name numerical control develop? Parsons had a contest at MIT to pick a name. Numerical control was chosen, and the winner received a $50 prize.
RO: What happened next?
Parsons: In March, 1952, MIT completed the NC machine, but they didn’t deliver the machine to me even though my contract with them called for delivery of it to me. They insisted that they could do a better demonstration than I could. That, of course, was ridiculous, because they didn’t know machining. MIT put on shows for the aircraft, machine tool, and electronics industries, and refused to invite me to attend those demonstrations. So I went to the Air Force and asked for invitations for Stulen and myself. The Air Force insisted that MIT invite me and Stulen. I was invited to all three demonstrations, and Stulen was only invited to one. At one of the demonstrations, a fellow from a Swiss company said to me, “you’re about as welcome around here as a bride’s mother on a honeymoon.”
At dinner after the MIT demonstrations, a Pratt & Whitney engineer said, “We don’t see anything patentable about this process.” I said, “I don’t agree,” and I went to work getting the patent. I managed to get those patents even though MIT was trying to patent the process at the same time. In January, 1954, I hired a patent law firm to do the work for 25% of all fees and royalties.
A patent was issued to John T. Parsons on January 14, 1958 three months before the MIT filing. The inventors were John T. Parsons and Frank Stulen. Patent No. 2,821,187.
RO: Once you received a patent, did you license it, or continue to develop NC yourself?
Parsons: After the initial MIT demonstration, and obtaining a patent, the next step was to find somebody to license NC. I worked very hard on getting a company to develop it. I decided to go to Bendix, because of their Detroit headquarters, and because it was easier than going to New York or out East to a machine-tool builder. The director of their research lab was A.C. Hall. I believe he was one of the founders of the servo lab at MIT. He had been there, and he understood the concept. He reported to Lawrence Highland, who was director of research at Bendix. I made several trips to Bendix over a period of four to six months, and then Highland called my patent lawyer, Richard Mason, and myself to Detroit for a meeting.
He said, “We’ve decided there is just nothing in the patented process that we can use.” I embarked on an 11-minute soliloquy and ended with, “If I had properly explained this thing to you, you could not have possibly made that decision.” He said, “if you feel that strongly about it I’ll send Hall down to MIT to have another look.” He came back, and we had a contract, an exclusive license to Bendix, with the right to sub-license. They were very dilatory in sub-licensing. They had granted some sub-licenses, but I figured there were at least 30 companies infringing my patents, and I couldn’t afford to fight them in court. Later, in 1970, I had a meeting with Bendix, and said, I can’t fuss around with the patents anymore. I want you to buy a paid-up license for one million dollars within 60 days, or else I will move my auditors in. Under the agreement with them, I had the right to audit their sales, on which I collected royalties. So I put the heat on them, and I got the contract and money. I had to pay Mason 25% of anything I made on the patent. Mason said, “John, you should give a percentage to each of four MIT engineers for their work.” I ended up giving 3% to the engineers and to MIT and 10% to Stulen in 1970.
Bendix was producing some machines themselves, and they were also granting licenses. To give you an idea, IBM took the biggest license, which was around $750,000. Fujitsu took one for $700,000 and GE for the same amount.
RO: Why did it take so long between licensing the patent and the widespread use of NC?
Parsons: The slow progress of computer development was part of the problem. In addition, the people who were trying to sell the idea didn’t really know manufacturing—they were computer people.
The NC concept was so strange to manufacturers, and so slow to catch on, that the US Army itself finally had to build 120 NC machines and lease them to various manufacturers to begin popularizing its use.
In March 1968, Parsons sold his company to a California conglomerate by the name of HITCO, and was to have stayed on as head of the Parsons division in Traverse City. Because of business differences, however he resigned six months later, and created a consulting and design firm to pursue development of technology with the US Navy for the manufacture of controllable-pitch propellers.
He had also begun work on the development of a milling machine to produce polystyrene patterns, and another machine he called ParTape, for automatic programming and tape preparation for an NC machine. It was a takeoff on the tracer mill concept, all intended to make numerical control more accessible in manufacturing.
The prototype machines were never completed because of problems with the vendor. After exhausting financial resources, including proceeds from the Bendix agreement, Parsons found himself in a situation not unfamiliar to many great innovators. He was rich in ideas, but had neither the money nor the clout to take them to fruition.
The Parsons’ Corp. had its own small helicopter to demonstrate its rotor blades. John Parsons is on the right.
As a consultant, he was involved in several projects throughout the 1970s dealing with alternative energy sources, including the conceptual design of a wind-energy system that would use a 420' diam blade to produce power.
A decade later his accomplishments began to be acknowledged outside the industry. In 1985 he and Frank Stulen were among the first recipients of the National Medal of Technology, which was presented to them by President Ronald Reagan. In 1993, Parsons was inducted into the National Inventors Hall of Fame.
Parsons personally received approximately 15 US patents in the fields of numerical control, marine propellers, foundry systems, and data-acquisition methods, and his business was awarded about 35 patents. He also pioneered adhesive bonding in metal aircraft structures, then built the first all-composite airplane. His technology revolutionized the production of conventional and controllable-pitch ship propellers. He produced fuel lines that were 20” in diameter and 22' long for the Saturn booster that launched the US astronauts to the moon. Parsons brought computers to aircraft design, manufacturing, and real-time management reporting. He developed evaporative patterns produced by NC to replace weldments with streamlined castings, which revolutionized the production of automobile body dies.
Although he has become partially blind from macular degeneration, at 87 Parsons is still very active. A resident of Traverse City, MI, he is currently working on ship-propeller designs, and is preparing an autobiography with Pat Green, a writer. Green says that Parsons was regarded as slightly eccentric in his business dealings. For many years he traveled from one appointment to another in a Greyhound bus (an MC 7), specially outfitted with worktables, a bedroom, and bath, with space for a Hammond organ, but he never slept in the coach. He held off on the organ, deciding that the $3500 expense was a bit much. Also, at one time he purchased five P-38 fighter planes even though he wasn’t a pilot and hated to fly. If you’re wondering what happened to the planes, and why Parsons bought them, you’ll just have to read the autobiography when it’s published in the near future.
According to information from an MIT source, the Cincinnati Hydro-Tel became the first NC machine. It was used for demonstration purposes at MIT. In 1959 it was offered to the Smithsonian, but they turned it down on the premise that two full-time people were needed for operation and maintenance, which they couldn’t fund. The system was put up for bid in 1962 by the government as surplus property. An individual purchased the electronics, and the machine tool was sold to the Cincinnati Milling Machine Company (now Cincinnati Machine), where it was reconditioned and resold.