Nuts & Volts - September 2004

In The Trenches

Of course, the transformer doesn't have to provide 2 amps. A 1.0 amp or 1.2 amp transformer could be considered. It will also be less expensive.

There are a number of points here. A conservative transformer choice (2 amps) makes other parts of the design non-conservative. The specification of an RMS voltage ( 24 VAC) is not the specification of a peak voltage ( 37. 3). Safety requires the full understanding of every detail. Cutting corners or carelessness is simply unacceptable when safety is involved.

It is also interesting to note the dance between target specifications and performance specifications. We make our design meet our target specs by choosing parts with proper performance specs. When we're done, we'll have a product with performance specifications. Further, our product may be used in another product to meet its target specs and so forth and so on.

Software Specifications

First, stop laughing. There is some truth in calling software specifications an oxymoron. If you don't believe this, read Microsoft's

"End User License Agreement." It is truly educational.

With software, if it works, then it meets specifications. There is no easy way to determine if standard engineering practice is employed. There is usually no outside review of the actual code. This is very different from hardware. A circuit board must not only work, but must meet specific design rules (AKA specifications) in regard to parts placement, layout, trace widths, trace spacing, hole size, and so forth.

There are no equivalent rules for software. There should be. There have been attempts to analyze code with software, but this has not been generally accepted, nor, as I understand it, is this software particularly good.

The target specifications for software become its performance specifications.

However, these are not the same. The target specs say what the product must do. The performance specs say how well it must do it. I have never seen actual performance specifications for software, with the exception of benchmark ratings. The only way to compare similar software packages is to actually try them, side-by-side.

Objective software standards are not impossible to create or measure. How about measuring the ratio of branch to in-line statements or a code-line to comment-line ratio? What about providing the number of standard library code-lines and the number of custom code-lines? These values would allow the comparison of one program to another. It seems to me that they would be useful to have.

Unfortunately, I don't see this happening soon.

The Root of All Legal

One of the most common sources of customer complaints and legal action is because the customer feels that the product doesn't meet specifications: This clock doesn't keep time. The radio doesn't pick up my favorite station. The air conditioning doesn't keep the whole office cool. I want my money back!

There will always be unhappy people. However, your product shouldn't make them that way. Properly specifying your product can go a long way in reducing these complaints.

Additionally, good specifications

· For the experimenter.

· Determine surplus stepper motor specs using simple easy to build test equipment.

· Design and build microcontroller-based control systems (flow charts and code examples included).

· Design and build stepper motor driver circuits.

· Analyze the mechanical characteristics of stepper motor-driven devices.

· The book is full of experiments, circuits and code. $34.95

ELECTRONICS

P.O. Box 1414, Hayden, ID 83835

Voice (208) 664-4115 Fax (208) 772-8236

We have been selling on the Internet since 1996. We ship the day we receive your order or the next business day.

SEPTEMBER 2004

New Titles

Same Useful Books Table Of Contents And Ordering Information On Web Site. Not Distributed Via Book Sellers We accept VISA, MC, AM, DS, MO, Check CA residents please add 7.25% CA sales tax See our web sites for s/h rates. PIC, PICmicro and MPLAB are trademarks of Microchip Technology Inc.