If you are not a ‘cat’ person, you might not appreciate the full advantages of computer assisted testing. I say ‘assisted’ because the bulk of the critical elements involved in testing remain in the preparation and setup of each test element. To date no one has found a way to ‘automate’ that. What has been accomplished is to take the repetitive drudgery out of the detail of equipment adjustment, data reading and data recording. Along with removing this drudgery, computer assisted testing has removed most of the conditions that, in the past, resulted in the introduction of human errors in the process.
My early experimentation with software control of testing began in1980 with a dual floppy drive desk top computer running Basic programs driving a rack of receivers (see accompanying photo) to make EMC emissions measurements. Previous to this, my technician and I would set up each test, tune each frequency range searching for a ‘peak’ in the meter (Yes, a meter with a dial and pointer!), write down the reading and proceed to the next range and repeat the process. To perform one scan from 10 KHz to 1 GHz by this method then plot the data consumed an eight hour day.
Our first scan with a single antenna with receivers under computer control produced a completed plot in less than 20 minutes. I recall my technician asking; “What do I do with the rest of my career?” With this quantum leap in test time, the design engineers realized that EMC was no longer just the final test their design had to pass. EMC was now a viable development tool. We began hiring technicians to keep up with the demand for test services.
The increased confidence from having the computer take the data, apply the antenna and cable loss factors, as well as store both the raw and the final data, removed much of our early debates with development engineers about “How confident are we in the data?” By careful use of test ‘master’ artifacts (comb generators) and the use of statistical process control (SPC) to track system stability, questions about the reliability of the test equipment have almost disappeared completely. The ability to have the system also take data and place it into a final report ensures that nothing is inadvertently altered or changed while it is being ‘handled’ by humans.
However, in my opinion, the greatest advantage with a well implemented computer assisted test system is the ability to free up the test technician to act as an intelligent pair of eyes, watching the test as it progresses and asking; “Do I believe this?” With EMC testing, more than any other, there is so much that can happen to fool us into believing that a computer drawn data chart ‘must be right’ when there is something fundamentally not right about the test. Our only defense against this is the computer between our ears, and a good sense of skepticism. A well developed computer aided test control system provides that freedom of action. 
 |
Kimball Williams
is a Technical Fellow for Denso Americas based in Southfield, Michigan, acting as the engineering lead for the EMC laboratory. He received his BSEE degree from Lawrence Technological University in Southfield, Michigan. Prior to joining Denso Americas, he was the Principal Designated Engineer for Underwriters Laboratories for 3 years. He began his EMC career in earnest as the Principal EMC Engineer for Eaton Corporation, where he remained for 26 years. He is a Past-President of the IEEE EMC Society and is presently serving on its Board of Directors. |
