First-Hand:Circuit Design, Fiber Optics, Games, Detector Arrays, Voice Communications: A Journal of an Electrical Engineer
Submitted By Harold Minuskin August 22, 2012
This article examines my electrical engineering experience, specifically focused on (but not limited to) the radical changes of circuit design from the 1960s to the early 1970s. I also relate how I designed electronic games before video/microprocessor games became popular. In addition, I recall my 22 years at a NASA site, rounding out my 50 plus years of experience.
The Beginning of My Career
After my third year at The College of the City of New York (CCNY), I started my electrical engineering career in 1959 with a summer job as a Junior Engineer for the New York City Transit Authority. I was provided a drafting table on which I prepared inked drawings to show how to convert passenger station illumination from incandescent to fluorescent lighting. In 1960, just prior to graduation, I took another summer job with Raytheon in Massachusetts. My assignment was to design test fixtures for vacuum tubes.
After graduation, I accepted a job with Douglas Aircraft Company. My first assignment was to design a circuit that would only allow every other pulse through. The output pulse would be used for timing and synchronization in the next stage of the circuit. The components on hand were pentode vacuum tubes along with the resistors and capacitors. One had to be very careful handling and connecting the high voltages required for plate and grid elements. I had already experienced several high voltage shocks when I was working in the lab during my senior year at college. That is how I discovered that a single layer of standard masking tape has the dielectric properties to insulate you from a shock up to 500V.
In the telemetry data processing center, where I first worked, I remember there were about 20 equipment racks, each of which contained multiple circuit modules. Each module was a flip-flop or similar type of circuit and was made up of at least two vacuum tubes mounted on a circuit board along with resistors and capacitors, and at times a fuse. The module was about three quarters the size of a masonry brick, and had a “pull to remove” handle.
I recall the enormous amount of heat generated by the equipment racks that contained these multiple vacuum tube circuit modules. A plenum floor provided cooling, in addition to fans and powerful air-conditioners that serviced the room where all the racks stood. It was an extremely noisy environment, which would never pass today’s OSHA standards.
Within six months, the use of vacuum tubes had deferred to germanium transistors. Within the next year, most of the circuit designs were using silicon transistors and diodes. At that time, the price of a silicon transistor was relatively high compared to a germanium transistor. However, the superior performance of the silicon components often justified paying the higher price. Amplifiers, active filters and other circuits, could now be designed without the need for 6.3V or 12.6V to power the filaments and the high voltages required for the plate and grids of vacuum tubes.
The cost of silicon transistors and diodes eventually came down to where such components were relatively inexpensive, and could even be purchased by hobbyists. About 1967 or 1968, a Sylvania electronic representative introduced me and several other electrical engineers to the first line of integrated circuits, called Sylvania Universal High Level Logic (SUHL). These integrated circuits provided logic elements such as OR, AND, NAND gates, as well as Flip Flops and other logic circuits in a very small package (about ¼ square inch).
The use of these logic elements was only limited by your imagination. A computer designed with these devices could now reduce its size and power consumption. It was now much easier for me to design digital logic circuits including test equipment. However, analog circuits still had to be designed with individual components. The data sheets for the SUHL integrated circuits are still available on the Internet.
About 1969-1970, circuit design engineers were invited to attend manufacturers’ seminars (actual sales pitches by Texan Instruments and other vendors) for a new breed of integrated circuits called TTL, transistor–transistor logic. We were given data and application sheets along with free samples of the “7400” series logic circuits. The trade magazines often displayed the use of 7400 series logic elements as analog design elements such as oscillators, inverters and amplifiers. I had two favorite uses for the inverter gates. One was to remove the “bounce” produced by a mechanical switch, or relay contacts. The other was using a series of inverter gates for the design of a low cost oscillator.
I worked on a project where a one-foot wide by two-foot high circuit board was populated with about several hundred of the 7400 series logic elements mounted on sockets. The reverse side of this large circuit board contained wire wrap posts for each pin of the sockets. A specific number of computer functions were designed by using wire wrap from one socket pin to another. This very large circuit board was mounted on hinges so that access to either side of the board was readily available. When this large circuit board was mounted in an equipment rack, fans would force cool air around the logic elements.
An advantage of the 7400 series components was their availability as the 5400 series for military applications with operational temperatures between -55°C and +125°C. The 7400 series was also available for the industrial temperature range of -40°C to 85°C. CMOS logic circuits emerged shortly after the 7400 circuits were introduced. The CMOS circuits had the same functions of the 7400 series, but required several orders of magnitude less power. For example, a single TTL gate circuit consumes about 10x10-3 watts, while an equivalent CMOS gate only requires 10x10-9 watts.
The introduction of the 709 and the 741 operational amplifiers was a huge leap for analog circuit design. When this device became readily available, it freed circuit designers from going through the rigors of amplifier design and analysis, component by component. You no longer needed to worry about temperature stability, oscillations, gain and other factors associated with amplifier design. The manufacturers furnished us with application notes and data sheets. In most cases, one could design a stable, highly reliable amplifier with just a few external capacitors and resistors. The LM324 amplifier became available about 1975. One eight-pin Integrated Circuit (IC) package contained such four amplifiers. In the early 1970s, the 723-voltage regulator and the NE555 timer/oscillator made circuit design tasks even easier. Now I could easily design reliable multiple supply voltages, timers and oscillators with just a few external resistors and capacitors connected to the small IC package.
A note is in order at this point about circuit design software programs. I had used such programs to analyze my circuit designs. However, I found it time very consuming just to learn the programs. Besides that, the programs were expensive; at times you had to model your own specific components because those devices did not appear in the program’s library.
Before Atari and the explosion of video/microprocessor games, I designed simple games using 7400 series logic circuits, NE555 timers, lights, and mechanical switches. I took on this new task as a consultant, in addition to my regular job. These games were eventually installed in bars, lounges and military recreational rooms. For one or two quarters you could play the game. In one such game, there were ten light bulbs that lit up, one at a time. Next to each light bulb were two rows of push-button switches, twenty switches total. The light bulbs were in the center. On each side of each light bulb was a push-button switch. The game began when a coin was dropped into a slot and one of the ten light bulbs would light up in random fashion.
Player A or Player B pushed the button on their side of the lit bulb. The logic circuits were designed so that the player who pushed the button first would earn a number of points. The number of points was determined by the location of each light bulb near the bull’s eye. Each light bulb was placed at various locations from a target. The closer the light was to the target’s bull’s eye, the more points a player would earn. The game continued as bulbs lit up randomly and each player tried to beat the others by pressing the appropriate button (near the light bulb) first. Some of the circuits I used were two sets of 74193 4-bit counters and two sets of RCA Nixie numerical displays. The player with the most points was the winner. A single player could also play the game. In that case, the player would try to earn as many points in the allocated time of the game, typically 120 or 240 seconds.
The NE555 timer was used to time out the game after about 260 seconds but could be adjusted from 60 to 360 seconds of playtime. The switch bounce from the coin activator would generate a random number of pulses into a 74193 4-bit counter, which would in turn illuminate at random one of the light bulbs.
I also designed a “bumper car” toy. This toy was the half the size of a masonry brick and formed in the shape of a car or animal. The toy contained two AA batteries, two micro-switches, some logic circuits, a transistor, and a small DC motor. When the device was turned on, the DC motor would drive the toy’s wheels in one direction until the toy would run into an obstruction. This would trigger the front micro-switch and would cause the motor to reverse direction. The toy would then run in the reverse direction until its back end micro-switch was activated and the drive motor would again reverse. The toy continued to operate until it was turned off, or the batteries ran out.
A variation of the “bumper car” toy was the “invisible leash toy.” It was designed so that a four to seven year old child could shine a small flashlight at a toy animal, such as a dog, cat, duck, etc., and the toy would come toward them, or follow the beam of light. This toy was about the same size as the “bumper car.” It contained two AA batteries, a small dc motor, a transistor, an HP light sensitive diode, and some other components. The light sensitive diode was recessed within the toy so that only a direct beam of light would activate it. This was an indoor toy for obvious reasons and generally functioned best in a low light environment.
My children, then in their pre-teens, would often serve as “beta testers” for the games and toys I designed.
My job would be completed once I delivered a working electronic prototype and a schematic of the toy or game. The electronics I designed would fit into an allocated cavity of the toy or game. Coin operated games required extensive artwork as well as stand-alone platforms. The outside of the toys required creative designs to attract the attention of the children. Once my customer completed the prototype toy or game, the product would be demonstrated to toy companies. If accepted, the design would be reevaluated for low-cost and large-scale manufacturing.
1980s and early 1990s
I was keeping pace with technology and designed stepper motor control circuits used in microfilm applications. This is where a camera is indexed a precise amount so its film is exposed to the alpha numeric data displayed on a Cathode Ray Tube (CRT). Multiple stepper motor indexing created about a hundred images on about 5 inches of 70 mm film. Each image can later be magnified to an 8-1/2”x11” page of information.
I also worked on a then-classified project involving secure fiber optic laser communications and fiber optic test instruments. With security of paramount importance, it was decided to encase the laser communication circuits in concrete to make them tamper proof. The fiber itself was made secure by constantly monitoring its transmitter power. Any deviation from a preset value would trigger an alarm. Fiber optics were also used to calibrate laser range finders commonly employed by tanks on the battlefield. A ranging laser emitted from a tank would be shot into a coil of optical fiber the size of a coffee can, which may represent a target at 500 or 1000 meters. The time of the return signal from the end of the fiber would indicate a specific target range. In this manner, a tank’s laser ranging system would be calibrated without exposing the laser to free air where it may harm personnel or damage equipment.
With the introduction of the 4004 and the 4040, 4-bit microprocessors and later the 6800, 8-bit microprocessor, I was able to design automatic test equipment for the wiring assemblies used by the automobile and airplane industries. To design programmable test equipment, I had to program in both Machine language and Unix. Initially relatively small machine language programs were placed on EPROMs to execute a program in a 6800 microprocessor controlled instrument. The advantage of the EPROMS was the ability to erase the existing program and write over another program. Once the test equipment program was debugged and ready to go into production, the EPROMS were substituted for the much less expensive write-only PROMS.
In the mid 1980s, desktop computers were being made available to engineers. When I worked at Hughes, there was a computer room with about five or six early Apple Mac 128K computers. One needed to wait for an available computer. You would bring your small floppy disc and work on your task. Mostly it was utilizing software programs for word processing, spreadsheets, or functional block diagrams.
Mid 1990s to 2011
Some of the interesting work that I performed in the mid 1990s dealt with 128 by 128 detector arrays, and later with 256 by 256 arrays. These arrays were cooled with liquid nitrogen to reduce the ambient noise. The output of the arrays was fed into a PC and special software would process the signal. The array was exposed to a telescope at the Table Mountain facility, outside of Wrightwood, California. The telescope would be pointed to view a planet or the moon. Various colored filters would be placed in front of the detector array. A dedicated camera would take a number of images corresponding the array’s output. In this manner, scientific data could be compiled about gases or other phenomenon emanating from the planet or moon. The only minor hardship I had to endure was to sleep during the day and work at night when the large telescope would be open to the sky. The upside of this month-long assignment was meeting with world-renowned astronomers who were the major users of these telescopes.
In the late 1990s, the potential Y2K problems appeared. Fortunately the organization where I was working at the time allocated sufficient time to review and fix any problems. In my case it was a matter of working with vendors who had sold us equipment. For most of this equipment, just a few lines of software code needed changing for Y2K compliance.
My knowledge of circuit design and hands-on hardware experience made me a good candidate for Lead Project Engineer for complex voice communication systems. These systems were needed to communicate between the various NASA sites, especially the three NASA antenna facilities located on three different continents.
Working at a NASA site, protocol dictated that any procured hardware or software over a certain dollar value be sent out for bid to at least three outside vendors. There were no shortages of vendors who vied for this $3-$5 million dollar contract. As Lead Project Engineer, I was responsible for the entire project from conception to deployment. Because of several requirement changes along the way, this project was initially delayed and a new schedule was prepared. As the design matured, hardware and software testing proceeded. In the end, we had a functioning state-of-the-art communication system that encompassed NASA sites, including the three antenna sites. The system functioned flawlessly for over 16 years before it was replaced with the next generation of communication hardware and software.
The most stressful part of my career was not about difficult designs, nor with meeting deadlines or budgets; it was when I was called upon as an expert witness in a class action lawsuit, which involved the possible failure of a car’s air bag initiation circuits. I was deposed for about eight hours. During that time the opposing attorney was determined to find a flaw in my circuit analysis of a car’s air bag deployment system. The testimony I provided, along with my out of court research and evidence, enhanced the class action case. The case was later settled out of court.
As I entered into my mid-60s, I decided to work half time. At this point in my career I was designing multiple racks with server arrays that would carry both data and voice traffic. In a server environment, reliability, Mean Time Before Failure (MTBF), Mean Time To Failure (MTTF), and Mean Time To Repair (MTTR) are extremely important factors. I enjoyed preparing reliability and availability models. My best reference was the book High Availability Network Fundamentals, by Chris Oggerino. I was able to calculate all of the aforementioned parameters based on the specific hardware in use. This data proved invaluable to show the high reliability and availability of the hardware and software that was required in a communication and data processing center.
In 2011, I retired at age 73.
Over my long career, I was very fortunate to have the opportunity to take on the challenges of many different engineering assignments. I was able to mentor younger engineers and established special relationships with my colleagues. I enjoyed what I was doing, and was even paid for it!