A HISTORICAL PERSPECTIVE INTO GROUNDING AND SHIELDING

As a teenager, in 1940, I built my own crystal set. I remember connecting the circuit common (ground) to a water pipe some 20 ft away from my set. The conductor went out through a hole in a wire mesh window screen. I found out that routing the antenna made a difference, so I kept trying different approaches to receive more radio stations. Little did I know of the complex nature of coupling to a transmitted radio signal. This was my first association with grounding (making an earth connection).

My interest in electronics extended to how a radio works and before long I was spending time in the neighborhood radio repair shop, where I learned how to test vacuum tubes. I was given a discarded radio as a present because the plastic case had been smashed. I freed the speaker cone and I had my own working radio. The radio had a ground clip that suggested that a grounding wire might improve reception. As a teenager, I spent time with several classmates that had acquired the skills needed to be amateur radio operators. They were always discussing their antennas and the grounding of their transmitters. I borrowed a copy of the ARRL handbook to get some idea of their hobby and what it meant to be a ham radio operator. I was an observer as I did not have the resources to enter this hobby.

I was drafted into the army in WW2 at age 18. Eventually, I became a radio repairman in the infantry and fixed radios in Patton's third army as it crossed Germany. The radios I serviced had no connections to earth as they had to be very mobile. I never gave grounding a second thought. After I returned home, the GI bill gave me the opportunity to go to Caltech and get a BS degree in physics. I remember taking courses in electricity and magnetism, not realizing the impact this subject would have on my future. I remember solving differential equations and fumbling through systems of units. I was introduced to Maxwell's equations. At the time, I had no way of assigning significance to this information. It was as if I was reading the first paragraph of many different chapters in many different books.

After graduation in 1949, I started working as an electronics engineer at a company called Applied Physics Corporation located in Pasadena, CA. My first boss was George W. Downs, a well-respected entrepreneur. During the war, he had worked as a high-level consultant and was associated with the Atomic Energy Commission. I had a lot to learn. The company products included oscillographs, electrometers, and spectrophotometers. I was impressed with the beautiful packaging and the fact they were so well respected by their customers. All of their products used vacuum tubes and I saw "grounding" for the first time. They explained to me how they used a grounding stud that collected all the common leads used in the instrument. This included the metal case, the equipment ground, the centertap on the secondary of the power transformer, the transformer shield, and the various circuit commons. There was no explanation given to me as to why this was the best solution. I was told that the order used in placing these conductors on the stud was important, and they had found a solution that made the instrument free of noise. In later years, this star-grounding configuration would appear in very unusual places. At the time, I had no basis to be critical of star-grounding methods. The products worked well and engineers with years of experience had spoken. Do not mistake me. A grounding stud was a valid approach to building this product. It is not however a solution to grounding in general. Asking questions did not yield useful answers and I did what everyone else did - I used common sense, I copied the procedures used in other products, and I experimented when I could. I was a part of the work force.

My first assignment as an engineer was to design a dc instrumentation amplifier. This type of instrument was needed in conditioning signals from strain gages, position sensors, and thermocouples. I was shown a circuit approach that had been developed by RCA that used a mechanical chopper to correct for dc drift. I was soon immersed in regulated dc power supplies, transformers, filaments, tube type selections, and feedback. I managed a design one channel of dc amplifier including a power supply that weighed over 70 lb. Do not forget that vacuum tubes take several hundred volts to operate and these voltages had to be very carefully regulated. When I look back at those early days, I can see how far electronic instrumentation has come and in particular how much I had to learn. At the beginning, there were no shielded transformers, feedback techniques were primitive, noise and hum were problems, and there was a limited understanding of signal isolation. There were selenium rectifiers that did not work very well. Dc amplifiers and vacuum tubes are a definite mismatch. In those days, that was all there was. The techniques of differential amplification and common-mode rejection had not yet entered my understanding. My boss was learning from me. We had to start somewhere.

The period after WW2 saw the growth of the aerospace industry. I was project engineer on several analog computers that were sold to Douglas, Northrup, and Lockheed. These computers helped in the design of the first commercial jet aircraft. The computer design was based on work done at Caltech and included some dc amplifiers I had developed. After this project was completed, our instrumentation group was sold to a company in the transformer business. Our first project was to develop a high-speed recording oscillograph. The photographic paper speed in this machine was over 200 ft/s. Needless to say, Kodak appreciated our business. Getting the paper up to speed in milliseconds was no small task.1 I designed the amplifiers that drove the galvanometers. I found out about the limitations imposed by using a common power supply to power a group of single-ended instruments. It was obvious that there was a lot to be gained by using a separate power supply for each signal channel. To meet this challenge, I began working on new techniques to reduce cost and size and avoid the use of common supplies. I invented a method of using AC coupling and a parallel feedback network to make a dc instrument. The company rejected my proposals for a new product line. I recognized the relevance of my new ideas, and I talked with two other engineers to leave and form a new company. George actually helped us make the transition.

The new company was called Dynamics Instrumentation. We manufactured instrumentation amplifiers for aerospace. The product line was based on the design ideas I had proposed. I could now contact users directly and I began to understand their dilemma. In rocket test stands, vacuum tube electronics had to be mounted in a blockhouse hundreds of feet away from the rocket engines and any sensors. This meant long input cables had to carry millivolt signals between structures. This raised issues of where to connect the input and output signal cable shields. I had some ideas on how to handle these issues and wrote some articles on the subject. I passed these articles out to potential customers. I was surprised at the reception these articles received. It was obvious there was very little information available on where to connect shields on large systems. Years later, engineers would pull these articles out of their files to show them to me. Now, when I look back at that period, I too had a lot to learn. I could tell that this was a difficult problem and the size of a company had nothing to do with understanding the issues. Rocketdyne was interested in designing rocket engines not where to connect shields.

I found out that interference resulted from the flow of power current in input conductors. Remember there were hundreds of volts on the secondary coils of the power transformers. Simply put, I was the culprit. This current could be limited by the use of transformer shields. I built my own power transformers and played with the shielding until I understood what was happening. My competition built a carrier-type differential dc amplifier that used a mechanical modulator/demodulator and a multishielded input transformer. Being differential allowed input and output commons to be grounded separately without creating a ground loop. I tried to duplicate this approach, but I had problems building the input transformers. Instead, I built a postmodulator/demodulator around a postcarrier transformer using newly available transistors and managed in effect to build a wide-band differential amplifier. The mechanical modulator approach had 100 Hz bandwidth and the post-transistor modulator instrument I built had 10 kHz bandwidth. I had a new product and I had a new definition of the word differential as applied to instrument amplifier.

I needed three shields in the power transformers I used in this design. I got a company in San Diego to build them for me. I noticed one day in an electronics magazine that this company was offering what it called "isolation" transformers with four shields. On my next visit, I asked the company owner what his recommendation was for using a fourth shield. He did not know. I then asked why he offered it. The answer was simple: "They sell better." I had aided in the formation of a new business based on adding multiple shields to distribution transformers. I had used shields to make one instrument work and the industry had decided to use these same methods to "clean up" systems. To me, they had a solution looking for a problem. To me, the multishield solution only worked for one instrument. Later, I would take a broader view of this shielding. I also saw...