Electronic Design

Bob's Mailbox

Dear Bob:
The degree vs. no-degree issue is more complicated than your readership realizes. There is great value to experience, willingness to attack a problem, and native intelligence and intuition. A person is born with some qualities, some are developed through education and training, and some can only be acquired through years of working in the real world.

The technical professions have identified a body of knowledge representing the core knowledge one must know to be proficient in that profession. The degree is a certificate indicating that the holder has acquired mastery of that core knowledge. It doesn't certify that the holder is creative, hardworking, or pleasant to be around.

The permutations of degrees, experience, native ability, and a dozen or more other qualities are what make the whole engineer. Holding a degree is only one of these qualities; it shoudn't be overrated, but on the other hand, it shouldn't be underrated either.

I have worked as a software engineer (programmer) for 23 years. I recently received a masters degree in Computer Science. The knowledge that I acquired through the masters course work would have helped relatively little in my career.

Conversely, I can now do certain kinds of analysis and development that I couldn't previously. I expect the same to be true of other technical professions.

There is no reason why a non-degreed engineer shouldn't take the time to complete a degree. Between testing out lower-level courses and credit for work experience (post fact, it's called an independent study class), it is not a large burden.

A technical course of study is focused and relevant to work. Yes, any good school will want you to jump through a few hoops, but the fact is, they will work with you to tailor a program if you work with them.

Assuming you have all the qualities that make a good professional engineer, why wouldn't you want to take the classes for a degree? You have to study continuously to keep up with new developments. So do your studying at a university where you can interact with a professor who has real expertise.

Your employer will help pay for the tuition and may even give you time off to attend classes. Time, money, value. Do you think you won't learn anything? You will certainly learn something you would not have learned just studying on your own.

With technology becoming more complex, and with the aerospace recession, certain job markets are more competitive. It may not be right, but the personnel people set standards that will prevent non-degreed individuals from being hired, and will lay off the non-degreed first.
Cory K. Hamasaki B.A, M.S., Alexandria, Va.

I sure agree with your first five paragraphs. But, there's "no reason" not to complete a degree? It's "not a large burden?" Hey, maybe I should steal the shoes off my baby's feet, ignore my family even more than I do, stop writing columns, and go back to school to learn how to design linear ICs from "a professor who has real expertise?"—RAP

Dear Bob:
Back in May, you put out a challenge for Fuzzy Logic applications. This is not one. In fact, all I have read indicates that F.L. designers have never heard of feedback circuits in control systems.

The examples of F.L. that I have read are one-dimensional systems: If the temperature is such, then the "comfort" parameter falls into cold/normal/hot to such and such a degree. A simple feedback controller can easily handle this kind of problem. Fuzzy Logic is overkill for an application that could be done more simply. Its only advantage is being able to make advertising claims that it's "Fuzzy Logic Controlled."

F.L. is an idiot level of Artificial Intelligence. AI can determine the correct identity or action from a complex input by using appropriate rules. The next higher level allows it to learn. The next level allows a "best guess" with incomplete data. For F.L. to shine, it must implement at least some of these principles. The degree to which a rule applies, and how it is applied, is determined by the data input. Rules modify the rules according to data. If data is missing or way out of line, fallback rules and constants may be used.

For example, F.L. that applies humidity as well as temperature would be a much more appropriate Fuzzy Logic application than a simple temperature controller. As humidity changes, the transitions between cold/comfortable/hot change position and possibly slope. At very-high humidity, the comfort zone may disappear. The controller would need to control both temperature and humidity through heaters, controllers, humidifiers, and dehumidifiers.

As long as F.L. is used in one-dimensional systems, it will flop. The more dimensions it uses, the better it will outperform other simpler feedback systems. However, the more dimensions it operates in, the more complex it becomes, and the more a true AI system may be appropriate.

I am not expert in AI or F.L., but rather in communications systems. But I do believe that F.L. will fail until it is used to perform judgment of appropriate behavior with incomplete or complex data.

I also agree that many articles are more F.T. (Fuzzy Thinking) than Fuzzy Logic. Thanks for the breath of fresh air on F.T. (or was that F.L.?).
Paul M. Schumacher, Ademco Inc., Syosset, N.Y.

Many demonstrations of new technology are greatly simplified. But most F.L. examples give no clear explanation of what works—or why! I hope my last two columns helped clarify that.—RAP

Dear Bob:
How I have waited to catch you in an error! The large Van de Graaff generators used at most physics research labs are horizontal-not vertical as you implied in the Oct. 1 column, "What's All This ESD Stuff, Anyhow?"

I can now die a happy man. (I might want to put it off a few years, though, since I am only 33.)
Chuck McCown, Data Com, Quincy, Ill.

Just like I said: Since most conveyor belts are vertical, those horizontal Van de Graaff generators are perpendicular.... I once thought I'd made an error-but that was a mistake....— RAP

Dear Bob:
I greatly enjoyed your column, "What's All This Ground Noise Stuff, Anyhow?" in the June 24 Analog Applications Special Issue. The importance of being able to make an accurate measurement can not be understated. I remember "Doc" Edgerton telling our circuit design class (circa 1955), "You guys will never be good engineers until you learn to take a measurement!" That admonition is probably truer today than it was back then, when you could almost get by measuring voltages with a wet finger and power levels with your sense of touch (or smell-depending on the degree of overload).

One statement you made in your article is more incorrect than not, and should be clarified. You say in reference to magnetic noises, "Unfortunately, copper or aluminum will not reject or attenuate these magnetic noises." If you check almost any book on EMI reduction techniques, you'll find references to attenuating ac magnetic fields with "reflective shielding" by using conductive enclosures made of those very materials.

An ac magnetic field induces a current in the shield, which creates an opposing field. I'm sure someone who remembers his/her electromagnetic field theory better than I do could derive a rigorous proof showing that ac magnetic fields are attenuated by conductive enclosures. I suspect the proof would show that even a dc magnetic field would not penetrate a completely closed box made of infinitely-conductive material. The material I've read states that reflective shielding is more effective than using magnetic materials above about 15 kHz, and the thickness of a reflective enclosure need only be about three times the skin depth of the material at the lowest frequency you're trying to attenuate. Holes for leads, adjustment tools, etc., should be kept as small as possible. Better one hole per wire than one big hole to take a whole bundle. The old "loop size" thing again.

Sometime last year I read a technical note from Hewlett-Packard that described the extremes they took to provide magnetic shielding for the front end of a wideband measuring instrument, with a multiple-layer shield that was constructed out of Mumetal for low frequencies and copper for higher frequencies.

I recently had to reduce the external 25-kHz magnetic fields produced by a small switching power supply in a spacecraft system. The 25-kHz H-field, as measured with a shielded loop antenna a few inches away from the supply, was about 20 dB over the allowable limit. I reduced these emissions over 30 dB by surrounding the supply with a cigarette-pack-size aluminum box with 1/16-in. thick walls and cover. To get maximum attenuation, I had to clamp down the cover with countless screws. An alternative would have been to use a metal mesh conductive gasket between the cover and the box, and fewer fasteners.
Jeffry A. Wisnia, MIT 57 EE, Jeffry Wisnia and Associates, Winchester, Mass.

Thanks for correcting my ignorance. Your comments are of great value—examples of problems in the real world, indicating that these very tricky principles are not well known. Many people read my Ground Noise column, but only you wrote in to educate us!—RAP

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.