If you arrived here from a search engine and did not find what you wanted, try the home page which has links to other pages at this site.

There are a few situations that a consultant cannot help. The first is to fool mother nature. I have noticed that physicists frequently fall into this category. They want a design that has performance parameters that cannot be economically done in the current state of technology.

The second situation is to want a flawed solution forced on a problem. Ham radio enthusiasts with no code licenses frequently fall onto this category. They have an oversimplified view of the problem and therefore think of a simple solution that does not fix the real problem.

The third is clients who do their own system engineering when they lack the skills. I have found that the more inflexible they are is insisting on their system design, the less talented they are in system engineering. The system engineering done by these people usually fall into two categories. Unnecessary complexity as an end in itself. This results in higher parts cost, higher manufacturing cost, and high labor cost to hand tweak each unit on the production line. The other fault is doing something different as an end in itself. Usually, the new way has been thought of before. The wise people who thought of it notices the pitfalls and saved their money. The foolish people tried it and lost big bucks.

The fourth is people who do not understand the skills needed by the consultant. They prefer someone who has done it before, even poorly, to someone who has the technical knowledge of the underlying scientific principles, has done projects that use these principles, and can produce an optimum design the first time. An alternative to this is to assume a time to complete the task and not ask the consultant for an estimate. A good consultant should be able to do the work in 0.5 to 0.1 of the time of a regular employee. Therefore, the hourly rate is not as important as the total task cost.

One semiconductor company wanted a consultant to develop application circuits for their products. They assumed that a person with large amounts of bench experience would be able to develop practical circuits. They chose someone who did such poor designs that it took months of mindless cut and try on the bench to make it sort of work most of the time. They turned down people who were good enough to have their designs work the first time and not need much bench tweaking. This is sort of like considering the child who repeats grades to be scholastically superior to the child who skips grades because the first child has spent more time in school.[Several years since writing this, the semiconductor company has failed to make a dent in the market with their new products that have been rejected by potential customers.]/P>