By Bruce E. Hofer
Prologue: When I was a senior at Franklin High School in SE Portland in 1965-66, my career aspirations profoundly changed when I took a shop course in electronics. Up until that point I had always thought I wanted to be a mathematician or mathematics professor. That class made me realize that electronics design engineering would be a far better choice. Indeed, it ignited a passion for performance and attention to detail that consumed me for the next 52 years of my life until my retirement in 2018.
I applied to Cal Tech in Pasadena during my senior year because I had scored a perfect 800 on the math portion of the SAT. I was accepted and received a scholarship offer. Unfortunately, even with the scholarship, the gap between the tuition and what I could afford was too great. Fortunately, I had also applied to Oregon State University where I was also accepted and given a very similar scholarship. I was still further encouraged to apply for a locally sponsored scholarship for science and engineering students. I no longer remember the name of the organization, but I was not selected. However, one of the review committee members felt so strongly about my situation that she arranged an interview with some people she knew at Tektronix.
That interview led to a temporary summer job offer working in an obscure support group in the basement of building 46. This group was responsible for designing, constructing, and maintaining CRT life test fixtures. The job paid better than my other options, so I accepted. I learned a lot about both vacuum tube and transistor circuit design from studying the schematics for those test fixtures. Ted Hunt, the head engineer of that group, took me under his wing and explained some of the finer points of those designs. That same group invited me back to work during the 2-week Christmas break and again the next summer. During that second summer (1967), I was introduced to some engineers and managers in the “ivory tower” (building 50). Their names escape me, but I now realize I was being screened for my potential as a future engineer.
During my sophomore year, Tektronix approached me with an offer I could not refuse. They had a summer student program in which they identified promising engineering candidates and offered to pay for the remainder of their education costs in return for a commitment to work for Tektronix for a period of at least 5 years following graduation. It was set up as a forgivable loan with a 5-year period of amortization. I quickly accepted.
At the end of my sophomore year in 1968 I was officially hired as a Tektronix employee and began to accumulate LOS. I was assigned to work in the engineering evaluation group managed by Leon Orchard on level 4 of building 50. One of my first projects was to investigate a problem with the battery pack in the 410 physiological monitor. The 410 contained a circuit that was supposed to automatically disconnect the batteries when their charge dropped below a certain value to prevent false readings (very important for a medical device!). But it tended to oscillate as the battery voltage dropped through the threshold causing the disconnect relay to buzz and sometimes fail shorted. I came up with a PUT (programmable unijunction transistor) circuit that neatly solved the problem.
One of my most vivid memories of that period was watching Russ Fillinger, then the project manager of the 410, getting upset with someone criticizing the product’s physical design. Russ took a 410, dropped it on the floor, and kicked it 20 feet down the hallway of building 50, then picked it up and proudly displayed it was still working! At that time the floors of building 50 were covered in smooth linoleum, not carpet.
During the next summer break between my junior and senior years (1969) I returned to work in the same evaluation group. I don’t remember everything that I worked on because that was also the summer of the Apollo 11 moon shot. The 7504 and 7704 and their various plugins were nearing production and there were many small and not-so-small problems to be addressed. I do recall reporting to Dick Mock and being involved in multiple component and circuit evaluations.
Following my graduation from OSU with a BSEE degree in June 1970, I was hired into the Advanced Products group managed by Hiro Moriyasu and reported to Carlo Infante. I was assigned to work with Bob Beville on the 7D11 digital delay plugin and given the task of developing a 500 MHz phase lock loop to be used in achieving the product’s goal of 2 ns jitter. That was a “sink or swim” project for a new, fledgling engineer with virtually no experience is such high frequency design. Carlo was not an easy person to work for, but I did succeed and eventually received my first patent with Bob Beville for the variable rate generator for the 7D11’s up/down counter.
It is important to note that the Vietnam war was raging during this time. I had managed to escape the draft with a student deferment while attending college, so I decided to extend it by enrolling in graduate school the next year. Thus, I attended OSU for a fifth year in 1970-71. During the summer of 1971, I again returned to work with Bob Beville on the 7D11. However, student deferments were abolished altogether around that time and the infamous draft lottery was instituted. My draft number was not high enough, so I was left with no choice but to enlist in the Oregon National Guard during the month of August. Indeed, I was sworn in only 3 days before receiving a draft notice in the mail (which I took great pleasure in burning). Although I joined the Guard in August, I was not scheduled to report for active duty until January 1972. That allowed me to continue my work on the 7D11 through mid-December 1971 and finish my portions of the design.
Unknown to many, Tektronix had a least one lawyer on staff who was busy applying delaying tactics for certain draft eligible employees who were very close to the age of 25. “25” was magic because the draft was not taking anyone over that age at that time. Because I had just turned 23 in the summer of 1971, there were simply too many months to go for me to run out the clock. So, I took a military leave of absence from January to July in 1972 for boot camp and occupational training. When I returned from active duty, I recall spending another couple of months helping Bob Beville finalize the 7D11 for engineering release. The 7D11 first appeared in the 1973 Tek catalog.
I was then hired by Gene Andrews and Les Larson in the Lab Scopes engineering group (headed by Oliver Dalton) and given the assignment to dig into a nasty production problem involving the 7904 and 7B92 time-base plug-in. Although one of the basic features of 7000-series was its supposedly interchangeable plug-ins, such was not the case with the original 7B92. My understanding is that production sometimes had to ship the 7904 mainframe and 7B92 time-base as a matched pair. Of course, this created logistic and servicing problems with customers owning multiple 7904 systems.
My first order of business was to build a test set-up to allow investigation of the sweep quality coming out of the 7B92 time-base. An older 7904 was modified to intercept the sweep outputs at interface pins A11 and B11, run them through a pair of bilateral constant resistance passive differentiators built by Les Larson (a story in itself), and then on to a pair of sampling inputs. The time derivative of a perfect ramp signal should be a pulse or rectangular waveform. Looking at the time derivative of the ramp would allow one to see startup aberrations and linearity in much the same way the vertical system can be evaluated using a fast step generator fixture, such as the 067-0587-02. The step generator fixture could not be used to test horizontal deflection amplifiers because of the relatively large capacitive load from the CRT deflection plates.
I was shocked by what I discovered--no two 7B92s were the same! They exhibited a wide range of sweep defects especially near the start of the sweep. In my humble opinion, the 7B92 was an example of the Miller integrator sweep generator technique taken too far. The fastest sweep rate of the 7B92 was 500 psec/div, 4 times faster than that of the next fastest time-base.
After trying my best to devise practical modifications to the existing 7B92 circuit, I was able to convince my superiors that the task was hopeless. I got permission to explore an alternate sweep generator design. Several months of fundamental research resulted in a new, clean method for gating and generating fast sweeps (see Patent 4,009,399). This directly led to the development of the Normalized Ramp Generator test fixture, the 067-0657-00 (the “nerge”) that allowed horizontal deflection amplifiers to be adjusted or “normalized” with much better consistency. Indeed, I received a congratulatory letter from Bill Walker who was then Vice President of Engineering.
Oliver Dalton was also very impressed that I had gotten to the bottom of the production problem, and he approved a proposal to develop the 7B92A based on my improved ramp generator circuit. Les Larson served as the project engineer and worked on the trigger circuits while I designed the sweep generator, delay pickoff, output amplifier, and assorted interface signals. The 7B92A first appeared in the 1976 catalog.
Around this time, I was also given some latitude in investigating the design of the horizontal deflection amplifier in the 7904. Horizontal deflection amplifiers can be quite challenging because signal amplitudes often exceed 100 Vpp at each deflection plate. One’s choice of transistors was limited by the high breakdown voltages required. Not sure who designed the original amplifier, but I discussed ideas with Joe Burger who had designed the then-current 2nd generation. Joe was very helpful and together we came up with some refinements that led to the 3rd generation design. I recall needing only 2 iterations of circuit board design before it was ready to be released. Between the 7B92A and the 3rd generation 7904 horizontal deflection amplifier, the production problem was over.
It was also around this time that I was asked to take over the instructor responsibilities of “AFTR”, Amplifier Frequency and Transient Response, an in-house course developed by Carl Battjes. I wasn’t sure I was up to the challenge, but soon became comfortable in this new role. AFTR was a virtual pre-requisite among analog designers. Carl’s focus had been solely on vertical amplifiers, so I added a lot of new material related to the design of other types of amplifiers. I ended up teaching the course for 7 years. Some years later when Carl became terminally ill, he begged me to “write the book” for him. Alas, I did not have the time, having moved on into a very different design realm.
My next project was to develop the sweep circuits for the 7B80 and 7B85 time-bases. These were to offer a fastest calibrated sweep rate of 1 ns/div with x10 magnification in support of the 7844 mainframe. I had further refined the gated ramp generator of the 7B92A to exhibit even smaller sweep startup aberrations, and vastly improved the “baseline stabilization” (the level of the sweep before it is triggered). I also developed a bootstrapped cascode output amplifier that effectively eliminated the annoying “thermal” shifts when switching the magnifier on. Tektronix also chose to introduce a 7B50A to replace the 7B50 to take advantage of the many improvements in sweep quality and triggering. The 7B50A was nothing more than a 7B80 with its fastest calibrated sweep rate limited to 5 ns/div via a modified cam-switch that locked out the 2 fastest settings.
Again, Les served as the project engineer and worked on the trigger circuits. I believe the 7B80 family were also among the first Tek products to use ICs instead of the more finicky tunnel diodes in the trigger circuits. The 7B80, 7B85, and 7B50A first appeared in the 1976 catalog.
It should also be noted that the 7B85 was also the first time-base to incorporate a feature called dual differential delay or “delta delay”. Instead of a single comparator to generate the delay pickoff trigger, it featured two comparators plus a digital voltmeter to measure the difference in pickoff voltages which translated to the delay time difference.
Next up came the 1 GHz scope system, the 7104. Using the test set I had perfected for looking into sweep linearity and performance, I was able to show that the 7B80 design could be adapted to provide a fastest calibrated (magnified) sweep of 200 psec/div. The sweep generator in the 7B10 family is almost identical to the 7B80 except for the addition of a second cascaded x2 magnifier to achieve an effective x20 magnifier for the fastest two sweep rates. At 200 psec/div, the entire screen represents a time window of only 2 ns. With the effective x20 magnifier, the viewable window extends to 40 ns allowing the triggering event to be easily found. The trigger circuits were designed by Art Metz whose exemplary work resulted in its incredibly stable triggering out to 1 GHz.
Having a little extra work bandwidth at the time, I was also assigned the design of the 7104 limited view-time feature. The micro-channel plate CRT used in the 7104 was very susceptible to permanent degradation (i.e., writing speed) if the beam current was too high or left on too long. The circuit ended up being a combination of beam current sensing and time domain integration against some predetermined limits. Not too challenging, but a very interesting project, nonetheless.
By the end of 1977 my work in Lab Scopes and time-base design had come to an end. My career interests took a dramatic turn from high-speed oscilloscopes and fast sweep generators to ultra-low distortion and noise audio signal generation and analysis.
Almost all analog design engineers have dabbled in audio amplifier design. During my early years at Tektronix, I fondly remember these often being referred to as “G” jobs. Many designs came and went with advances in high power semiconductor availability, especially high power PNPs then power MOSFETs. What particularly fascinated me was how to test audio amplifier performance. Almost by accident I met Bob Metzler, marketing manager of TM500 products. TM500 was considered a bit of a pariah group within Tektronix headed by Jerry Shannon. Bob Metzler and I had the same common interest in audio testing because he used audio test gear extensively in one of his previous jobs working in the radio industry. I was invited to join TM500 and begin development of a low distortion audio generator and corresponding distortion analyzer. Many of my 7000-series associates thought I had a hole in my head when I announced my transfer, but somehow I knew it was the right move for me.
Bob Metzler and I hit it off from the very beginning. Two other engineers in TM500 very quickly jumped at the opportunity to join me, Fred Armentrout (now deceased from a brain tumor) and Rush Hood (later left Tektronix to join Critikon in Florida). I was promoted to group manager during a major reorganization of the engineering department. About this same time, a recently graduate of RPI named Dr. Rich Cabot was applying to Tektronix for a position. His interest in the TM500 modular concept plus some product proposals he had prepared caused me to hire him on the spot. Three other engineers were also assigned to my “Low Frequency Products” group working on power supplies.
Our efforts resulted in the SG505 low distortion oscillator introduced in late 1979 and the AA501 distortion analyzer introduced in 1981. Both were state-of-the-art products at that time, and Tektronix soon became a household name in the audio test segment. We turned a lot of heads during those years because the AA501 was also the world’s first fully automatic distortion analyzer with auto-ranging, auto-tuning, and auto-nulling all of which were tedious manual operations with other analyzers. Indeed, even Howard Vollum took a personal interest in our products and he visited me on two separate occasions lasting over an hour each. I consider those meetings to be among the high points of my time at Tektronix. Howard died only a few short years later.
In the early 1980s Tektronix management seemed to be going through a phase change. Many of us called it the era of the MBAs or the era of the bean counters, but suddenly every group was going through exercises to justify their existence. Left on its own, I am not so sure TM500 would have survived. Then out of nowhere came the announcement that TM500 and SPS (signal processing systems) were being merged. Jim Cavoretto was to be the BUGM (business unit general manager) of this new entity. Jim and I knew each other well because we sat only 4 benches apart in Hiro Moriyasu’s group years earlier. Jim was working on another project while Bob Beville and I were working on the 7D11.
We were given the tentative mission to develop a series of programmable products to address an increasingly automated world. Because Jim knew my background, he made me chairman of a “Packaging” committee and charged my committee with evaluating the then-new corporate package developed by Marlowe Butler to house our new family of products. Other committees were formed to investigate market segmentation, manufacturing, profitability, and all the other aspects necessary to develop a good business plan. I never could grasp the merger of our two groups, but I was interested if it would lead to the development of programmable versions of our SG505 and AA501.
After about a year of this various committee activity, a consensus evolved leading to the TM5000 concept. This was to be modular ala TM500, but also with a digital interface (GPIB) to provide programmability. New products quickly conceived were the DM5010, FG5010, DC5009, DC5010, PS5004, PS5010, and CG5001 (a triple-wide scope-cal module). Bob Metzler and I fought an uphill battle with management, but we succeeded in getting the SG5010 audio signal generator and AA5001 distortion analyzer added to the TM5000 lineup. I was also put in charge of the development of the TM5006 (full rack) and TM5003 (1/2 rack) mainframes to power and cool the new plugin modules. The new mainframes also had to accept as many of the non-programmable TM500 set of plugins as possible.
I think most of the TM5000 products first appeared in the 1983 catalog; however, the AA5001 and SG5010 were completed a bit later due to my involvement with the TM5000 mainframe development. They were delayed until the 1984 catalog. My audio group was subsequently informed that it was to be disbanded. This came as no surprise to Bob Metzler and me because we had faced intense management pushback regarding the SG5010 and AA5001. With the writing clearly on the wall, Bob and I started having private conversations in the evening over beer or cocktails regarding what we should do. We simply had too much fun developing the audio test products as a team. We considered approaching TV Products in the Communications Division because of the obvious synergy between audio and video in the world of TV. But we chose a different path. Bob resigned from Tektronix in June of 1984 along with Rich Cabot and Bob Wright, two of the key engineers working on the SG5010 and AA5001 Wright. I resigned a couple of months later in August. My decision to leave Tektronix was a bit more difficult because I had two young daughters ages 4 and 6 at that time, and my wife had stopped working to be a stay-at-home mother. Although we would be facing some lean times while starting up a new company, my wife strongly supported my decision. On October 28, 1984, Audio Precision, Inc. was born.
In the 16 years I worked for Tektronix, from 1968-1984, I had designed major portions or influenced the design of over 20 products and received 9 patents along the way.
Epilogue: Following Tektronix, I spent the next 34+ years at Audio Precision developing advanced audio test and measurement equipment. Audio Precision was not a competitor with Tektronix, and we enjoyed good relations. Indeed, I was contracted by the TV Products group to design the audio board inside their VM700 product! In 2002 I became Chairman of Audio Precision’s Board of Directors and its majority shareholder. I retired in April 2018 after selling my shares back to the company. But all that is another story...