|
Departments |
Design Idea #1879 (EDN, July 4, 1996, pg 95) is a clever little piece of code, but it has omissions.
Instead of #define HasOne1(value)((value ((value) & (value–1))), as given, the correct form should be #define HasOne1(v)((v&&!(v & (v–1)))). Several logical symbols and parentheses are missing.
Everett M Greene
via the Internet
As a technician turned engineer, I read Jack Ganssle's "Do you need a degree?" (EDN, Aug 15, 1996, pg 147) with great interest. Ganssle's view is, at the least, very enlightened and unbiased. From the perspective of a manager looking to hire an engineer without a degree, Ganssle's view is accurate and correct. But, for any person looking to start a career in electronics, getting a degree should be your first choice. My experience makes the point.
I struggled for many years to find an employer that would give me the opportunity to prove that I could grow into an engineering position. I had a patent and 10 years of experience—three in production (which I hated) and seven as an engineering technician—before I found someone willing to let me show my stuff. The engineer I worked with at the time left the company, and my manager let me complete the engineering work on our project, even though I was a technician. After two more years of completing several more design projects and adding C programming to my skills, I was given an engineering title and put on a salary, which resulted in a pay cut. (I stopped getting paid for overtime and I worked more hours, which is typical when technicians go on salary.) My career was further enhanced when the company I worked for was bought by Hewlett-Packard, and I started being paid like an engineer.
If I had spent four years getting a degree and gained three years of work experience, five years of my life would have been spared getting to my current station. My trek to engineer was probably not the fastest or slowest in history. Of the large number of engineers I work with, there is only one other person who became an engineer without a degree. The other technicians I knew either went to night school to get a degree or had to quit to have their roles change. Some even got out of electronics entirely, due to frustration.
Based on the experience of this technician-turned-engineer, the fastest route to a successful career in engineering is with a degree. It doesn't mean you will be a better or worse engineer; it just means you can get there faster. For those of you well on the way, don't give up. It can and will happen to you if you want it and ask for it.
Teno Cipri
Manufacturing Development
Engineer
Hewlett-Packard
The functionality of current key-entry devices is still severely limiting and seems stagnant. When will the volume market offer plug-and-playable color flat screens for table or laptop? These screens should have touch-sensitive, transparent, exchangeable overlays for drafting/painting/typing and programming. They would also allow easy end-user composition of custom sets of selected characters, fonts, and controls and easy end-user construction and interchange of individually well-fitting, freeform layouts of these sets. The screens are widely needed now!
Rolf A Evren
EVACO
Danderyd, Sweden
I just finished reading Clive "Max" Maxfield's "Have you ever had one of those days?" (EDN, Sept 2, 1996, pg 111). Reading Maxfield's approach to the design brought me back to my college days, and it also brought me back to my first job. I worked for a Russian engineer who was in love with the Johnson counter. Because only one flip-flop changes state per clock cycle, the flip-flop outputs could be decoded "glitch-free" using combinatorial logic. The drawback is that you get only 23n states for n flip-flops, and a binary counter gives 2n states.
You can construct a Johnson counter by connecting all clock inputs to the main clock. For example, connect Q0 to D1, Q1 to D2, Q2 to D3…and Qn/ to D0. Connect the resets to Power-on reset. The counting sequence is 000, 100, 110, 111, 011, 001, and so on. In short, you can construct a three-stage Johnson counter and use the three flip-flop Q outputs to get your 608-out-of-phase signals. (Buffer them if you're driving heavy loads or long lines.) There you have it, nine minutes saved, also a few gates, and your cup of java didn't even get cold. Now about that 68…
Mark C Atchison
Standard Microsystems Corp
Hauppauge, NY
In "Low-voltage differential signaling yields megatransfers per second with milliwatts of power" (Sept 2, 1996, pg 119), equations 3 and 4, pg 120) were printed incorrectly due to a proofreading error. The correct equations are as follows:
VBKWD=(VA/TR)(L/2TL)(CCZ+LC/Z),
and
VFRWD=(VA/TR)(L/2)(CCZ–LC/Z).
The telephone number for Boulder Creek Engineering was incorrectly listed in the July 4, 1996, issue of EDN (pg 161). The correct number is (408) 867-8170.
Re: "If a little's good…," (EDN, Sept 26, 1996, pg 11): I call this the Hot Sauce Principle: "If a little is good, it does not follow that more is better." Unfortunately, the people that generate paperwork requirements haven't found this out.
I remember reading that Bill Gates had predicted about 10 to 12 years ago that computer capabilities would increase by a factor of 100 in 10 years and that we would see large increases in productivity by that time. Well, the 10 years have gone by, and the expected increases in computer performance have happened, but the productivity hasn't. The additional capabilities have not gone into productive work; they have gone into documenting compliance with regulations. Unfortunately, many of the regulators check the paperwork and don't check the work behind the paper. Is it any wonder that so many people don't believe their government?
Phil Spray
via the Internet
| EDN Access | feedback | subscribe to EDN! | Table of Contents |