Advertisement

Blog

The Engineering Desk-to-Bench Ratio

In companies I have been to, engineers usually are given both a desk and a workbench. The desk is where plan design and non-technical overhead is performed. It is the realm of, and symbolizes, theoretical activity . The bench is where prototypes are built and made to work — the symbol of experimental or practical activity. The bench has measurement equipment, circuit construction tools, and parts inventory is not far away.

The desk-to-bench ratio can be summarized as follows:

When I was at Tektronix, the desk and bench were two sides of the same U-bench, thus uniting the two kinds of activities into an integrated environment. The U-benches were 6 or 8 feet in depth, and the older, more established engineers had 8 foot benches. One side defaulted to a desk while the other was the bench. In the early mainframe-computer days, engineers who simulated circuits on computer had to go to a computer room where terminals were available for such activity. In time, the desks were populated with computers and the environment was a complete unit.

In the second and last company I worked at – a company started by a lawyer, not an engineer – the desks were all in an office environment and the lab was some distance away. This induced a kind of occupational schizophrenia, so in time I moved to the spacious lab and rarely visited my desk. This had the added advantage that few were in the lab and interaction with others in the form of distractions was reduced.

At Tek, Paul was one of my fellow college-student friends who went to MIT during the school year. He spent what seemed like inordinate amounts of time thinking through a design plan before he ever started building it. For resistors with values that were broadly optimal, he would select them so that their color codes had pleasing color combinations when next to each other on the circuit-board. Yet for an undergrad student, he was rigorous. Tek would involve “summer students” in design work and although it was not expected that they would be as productive as full-time, fully-qualified engineers, productive work did ensue from them.

Remembering those early days, I always intended to write a book (and recently did; Transistor Amplifiers at www.innovatia.com) that assumes only pre-calculus mathematics yet develops s-domain circuit theory (bypassing the Laplace transform) so that zealous, rigorous high-school students can become engineers with high-school math. Jim Williams of Linear Technology Corp. took psychology in college and probably did not think in the s-domain. His thinking was experimentally-oriented and done at the bench. Some good engineers operate this way, but in the end, theoretical depth has advantages if it can be related to actual circuits. The best engineers have depth across the full spectrum, from desk to bench. A few of them, however, operate like Jim Williams did.

Some engineers think that with enough intuitive circuits understanding and insight, it is possible to bypass all that math and still achieve the substantial designs. They do this by developing keen qualitative insights to design based on what in math would be the polarity of derivatives. Change this resistor value a little, and the gain changes hugely, or not at all. With a qualitative database of knowledge, even measurement instruments can be designed. When I was still in my teens at Tek, one of the business-unit managers, Jerry Shannon, liked to help technicians struggling upward. He would go over to final test and calibration and take some of the better technicians into engineering. Some of them worked out passably, but there were times of excitement. Roger S. was one of them. He designed the TM500 FG502 10 MHz function generator. It went into pilot production, but it had problems with the sine shaper. In an intensive burst, he put 71 modifications into the design – a huge number for a product already sent to production. Jim G. was another technician, an older, even-keeled guy with more experience. He ran into trouble with the DM502 DMM in the usual place DMMs run into trouble: around zero-scale. I remember that the modifications, which were one or two – not 71 – were based on intuitive, qualitative reasoning, sufficiently in-depth to achieve a workable design. As I look back at Tek circuit diagrams from that era of the ‘60s and ‘70s, it is amazing that so much equipment others depended on for their own design work was designed with intuitively-placed patches and ad hoc modifications.

In the design groups at Tek that produced the highest-performance products, there were people with significant depth at both theory and practice. One that comes to mind is Barrie Gilbert, still going strong in his 70s at Analog Devices. Barrie is an IEEE fellow and has a good intuitive sense of circuitry at the bench. Yet if you read some of his IEEE papers, he has more theoretical depth than many engineers. The oscilloscope wideband vertical amplifier designers also fell into this category – people like Carl Battjes, Thor Hallen, Val Garutz, John Addis, Bob Ross, Art Metz and Jim Woo. And for the fastest ‘scopes, the time-base was also a challenge to which Bruce Hofer, later the analog founder of Audio Precision, Inc. brought penetratingly new circuit ideas, one of which I used for the time-base design of a (slower) portable ‘scope.

My desk/bench ratio has been increasing with time, and I am becoming more like Paul Magerl. It is easier to change a circuit in a computer file or on paper than on a prototype board, especially if the change is extensive. However, if you are a younger engineer, do not expect to be as proficient at this as you will be later if you keep learning electronics. A lower desk/bench ratio is expected of less experienced engineers. As you work with actual circuits, you encounter some of the subtleties that are deeper than you have descended to at present. These subtler, unexpected behaviors of circuits can at first be frustrating, even overwhelming, as though the gods are being fickle in capriciously tweaking the physical laws as you try to master them. But don’t give up; the physical universe has a rational and trustworthy underpinning. It all makes simple sense when the causes of unwanted behaviors are discovered. Persistence can turn a feeling of resignation into one of curiosity, in how the circuit could possibly behave differently than you really think it should. This leads to a hunt for causes in an exploratory frame of mind. It is a quest for discovery of false assumptions in thinking about the circuit. And this leads to fabulous new insights, one of the personal rewards we as engineers are privileged to experience.

The perplexities arising out of bench work are a driver for deeper study into electronics. Sometimes a sophomoric view of engineering can develop which arrests further progress in gaining proficiency. This is the attitude that familiarity is the same as understanding. (My “Seemingly Simple Circuits” series on the Planet Analog website is in part intended to dispel this obstructive mental habit.) Circuits that become familiar often – almost always – have deeper subtleties that are not obvious to the novice who does not know that they are even there. Do not assume that all the important concepts – even basic ones that have continual application – are covered in formal education. A graduate degree in E.E. is no guarantee that you’ve descended below SCUBA depth.

As years of improvement as an engineer occur, you reach a point where not much that is new comes along — though this is never nothing! The more you think you’ve mastered electronics, the keener you are to find what you have missed. An engineer friend of mine who worked at Keithley on precision DMMs, Gary Bergstrom, sometimes ends his emails to me about technical matters with “What have I missed?” At this level of proficiency, the desk/bench ratio increases and maximizes productivity. It is quite satisfying to achieve the ability to design a complicated circuit or system, take the time to think it through carefully, and then build it and have it work almost, if not exactly, as planned. Eventually, the desk/bench ratio is necessarily maximized as failing eyesight and weakening motor skills leave one in the position of doing solely theoretical work.

14 comments on “The Engineering Desk-to-Bench Ratio

  1. Scott Elder
    November 12, 2014

    Hi Dennis,

    Nice succinct way to discuss engineering approaches.

    Having spent the majority of my career in the integrated domain, my desk/bench ratio has to be reported using logarithms!

    One twist on your blog is that we who work in the integrated domain don't have the luxury of “checking things out”.  Checking out a component value can cost $100,000 and potentially months of schedule time.  So most IC designers don't have much choice in the matter.

    That's also the reason that it is tough to land a job right out of school doing IC design.  One really needs that < <1 ratio at the beginning of a career.  I know it sure helped me.

  2. GSKrasle
    November 14, 2014

    Dennis,

     In a flippant way, I sometimes pretend frustration that 'I have yet to find an opportunity to write a paper overturning the Known Laws of Physics, because every design misbehaviour, every one SO-FAR, has come-down to something understandable, in-principle predictable, and often just a 'D'oh!' no matter how confusing, mysterious and frustrating at first apprehension.'

    By this I intend to impress upon myself, and my audience, the precedent that diligence and contemplation (and research) have always eventually provided solutions, experience/intuition and theory being, of course, salutary.

    I also say to myself, when facing a dauntingly complex or unfamiliar new assignment: 'it may be confusing, incomprehensible, right now to me, but SOMEBODY has understood it or does now, and so it is within my grasp should I apply enough diligence.'

    Despair never solved a problem, passed a test, completed a contract or put victuals on the table.

  3. D Feucht
    November 14, 2014

    Scott,

    Thanks for your example of where desk/bench > 1 is a necessity, in IC design. At Tek, in the rubylith days, designers would hang their IC layouts on the walls of the engineering building and others would stand around, looking for flaws in them, with some token reward if/when found. I remember in particular that George Wilson, inventor of the Wilson current mirror, would do this, and most of my discussion with him was in front of an IC layout on the wall!

    GS,

    Thanks for the encouragement and advice to those bogged down with seemingly unsolvable circuits situations. One of the reasons many of us became engineers instead of politicians is that we prefer to work with the most reliable Administrator of the universe versus those on the other end of that spectrum. As Henry Kissinger once said, “Corrupt poiliticians make the other 10 % look bad.” Happily, we do not face that problem in our work on the bench.

  4. Davidled
    November 16, 2014

    If ratio is equal to 1, then engineers might interpret themselves that they did both theoretical and practical way for any assignment. Today, computer or any kind of tool will support both two characteristics, which is important for engineer, since any engineering tool is so much advanced. I guess that some case might not fit to this formulation because engineer might face so many variety cases.  Ratio could be a baseline for engineer daily work.

  5. Myled
    November 18, 2014

    “In companies I have been to, engineers usually are given both a desk and a workbench. The desk  is where plan design and non-technical overhead is performed. It is the realm of, and symbolizes, theoretical activity . The bench  is where prototypes are built and made to work — the symbol of experimental  or practical  activity. The bench has measurement equipment, circuit construction tools, and parts inventory is not far away.”

    Dennis, in some of the companies employees are rotating between desk and work bench. Now a day's managers are good enough to utilize the employees working times. So for off project time they use to make use employees skills for bench projects (off shelf projects). 

  6. aklompe
    November 18, 2014

    “One of the reasons many of us became engineers instead of politicians is that we prefer to work with the most reliable Administrator of the universe versus those on the other end of that spectrum.”

     

    That is a very nice way to say you cannot fool nature. It's is this reliability that makes it possible to discover the laws of nature during a persons life time. Human nature lacks this reliability, maybe that's why we never seem to be able to solve big problems in the world around us, simple because no one gets to the expert level required to do so.

  7. D Feucht
    November 18, 2014

    “… we never seem to be able to solve big problems in the world around us, simple because no one gets to the expert level required to do so.”

    I don't think it is a lack of expertise that keeps social problems from being solved but rather a faulty exercise of the will by too many of us, starting with those at the top of the social order.

  8. Myled
    November 20, 2014

    “That is a very nice way to say you cannot fool nature. It's is this reliability that makes it possible to discover the laws of nature during a persons life time.”

    Aklompe, that's true. Nobody can fool nature and it has its own rhythm. 

  9. geek
    November 26, 2014

    “One of the reasons many of us became engineers instead of politicians is that we prefer to work with the most reliable”

    @D Feucht: I think that's certainly a trait common in engineers that they want to work with concrete ideas and proven facts rather than abstract debates (which politicians are somewhat good at). I think most people who end up becoming engineers and are succesful in their careers have had such traits since very early in their lives.

  10. geek
    November 26, 2014

    “That's also the reason that it is tough to land a job right out of school doing IC design.  One really needs that < <1 ratio at the beginning of a career.  I know it sure helped me."

    @Scott: That's exactly the advice I give out to fresh graduates these days that they need to get their hands dirty rather than look for the comfort behind the desks. Unfortunately, the culture these days encourage jobs which are more stable and don't require a lot of hands-on tasks. 

  11. geek
    November 26, 2014

    @Dennis: Isn't it normal for people to progress in their careers from bench to desks? Most people heading engineering companies these days are doing the desk roles but almost all of them have solid hands on experience working on the benches and only with time they managed to move up and got itno desk jobs.

  12. D Feucht
    November 26, 2014

    tzubair:

    I used  the word “desk” as a metaphor of theoretical electronics understanding or activity. As you noted, desk jobs, however, often do not depend on such ability and involve no bench work at all. They lie outside the metaphor yet are not uncommon.

    To the extent that technical leadership lacks technical understanding, projects associated with them are likely to fail. The most productive engineering group is one led by the most technically capable person. This seems like a truism, yet is often not the case in practice.

     

  13. geek
    November 29, 2014

    “The most productive engineering group is one led by the most technically capable person. This seems like a truism, yet is often not the case in practice.”

    @D Feucht: I don't think this can be challenged. I certainly think that it does require a solid technical background to lead an engineering team. You do need other soft skills like leadership team management etc on top of it, but more than anything else you need to have a solid technical know-how.

  14. D Feucht
    December 5, 2014

    tzubair:

    “… more than anything else you need to have a solid technical know-how.”

    I would add that a leader with outstanding technical capabilities yet with some less-refined social aspects can still be admired and followed by those who respect technical expertise. It is the person who is lacking in both technical and social skills that can be (to put it in German) ungenugent – deficient for the role.

    Richard Feynman made some comments about this once. He said that a person who is arrogant and also a master of what he is saying can be tolerated. But a person who is both arrogant and faking it, Feynman said, “that I cannot stand!” I agree with Feynman on this point.

Leave a Reply