In the beginning of a new-product lifecycle, a request for quotation (RFQ) is issued. Potential bidders examine the RFQ, see a section on EMC requirements, and decide to use an outside test vendor for support. A Web search results in many excellent laboratories specializing in EMC.
The bidder sends an RFQ to potential laboratories. Several of the labs request information about the product, but since the product isn’t designed, the answers are somewhat sketchy. One of the labs that provides a quote is awarded the job. The product developer then puts the unit together, ordinarily not involving the test lab until the next critical time.
Fast-forward several months and the lab is contacted to schedule testing as soon as possible but not later than next week. Sound familiar? Well, it’s not an unusual scenario. In fact, it’s so prevalent that I’ve compiled some of the questions most often asked along with some realistic answers to help you prepare for your new-product test to MIL-STD-461F.
Are you ready for test? is a question seldom posed until the schedule demands that the testing shall now begin. If the planning did not involve the lab, then you definitely are not ready to test.
But the testing proceeds anyway because the schedule demands it. I’ve frequently said that the qualification test portion of the development cycle is the time allocated to compensate for all schedule delays that have occurred prior to test.
Engineers designing the product sometimes are unaware of EMC requirements and, if they are aware, may face challenges implementing measures to control EMI. This calls for getting EMC design assistance or adopting the almost universal attitude of test and see what happens.
History has repeatedly shown that designing for EMC at the front end of the development yields a very high compliance acceptance or only minor tweaks to resolve the odd issue. Associated with many DoD-funded contracts is a requirement to submit an EMC control procedure, a documented design approach to implement EMC control measures. This document has a lot of value when provided to the engineers as a guide. To be really effective, it should be shared by all the groups on the design team: mechanical, electrical, layout, manufacturing, and test.
An EMC lab quotes the testing effort, and in that process, the required tests should have been identified. Normally, this is less than adequate instruction to perform the testing in a manner that will assure acceptance by the appropriate agency.
Upon arrival at the lab, the primary goal is to proceed with test regardless of the preparation: Time is money, and the delay hurts both the manufacturer paying for lab time and the lab because delays interfere with other projects.
The absence of a plan typically consumes a day or more of lab time just to establish get-by parameters for a chance that the testing will be accepted. The selection of correct limits and the applicability of test methods, test frequency ranges, and other elements are just a few of the items that the test plan/procedure should have addressed.
MIL-STD-461F requires that the same type of cables used for installation be used for testing and specifies lengths for qualification testing. It does not allow the use of shielded power cables.
So what happens when the test article arrives at the lab and the cables are incorrect? The first day of testing is not the time to realize that the cables won’t support the test installation.
The test plan/procedure should clearly define the complete wiring to be used for test, including type, length, terminations, and arrangement. Has anyone asked if the cabling arrangement will fit the laboratory constraints? Has penetration of the shielded enclosure been considered in a manner that will control ambient coupling?
Preparation and coordination with the test laboratory are essential to getting the test started on time. Having to remake the cables just to get started puts a severe delay in the process. It’s not out of the realm of experience to have products arrive at the lab with absolutely no cables and no discussion with the lab about configuring the UUT.
Have you asked the test technician monitoring the UUT during susceptibility test what constitutes a failure only to receive a blank stare? Do you know the acceptance criteria? Is monitoring available for measuring the acceptance levels? Is the monitoring equipment protected? Are the acceptance criteria quantifiable and measurable?
These seem such simple issues, but I have watched acceptance requirements change during the test, especially when a failure occurs. That usually means that the real criteria weren’t really known.
Sometimes the pass/fail criteria are not obvious, so communications between the lab and customer representative are critical. For example, I recall a UUT with a temperature-controlled fan that activated when the unit was hot. No lab personnel were even aware of the existence of the fan until the unit failed—permanently. Seems that the radiated susceptibility signal caused the temperature sensor to register a -128°, and the fan never operated. The warm enclosure and lack of monitoring the fan allowed the UUT to overheat and fail.
Not every lab has 115-VAC/400-Hz/75-A service available at the wall, and locating a generator to rent with that capacity can be a significant task. How was this special flavor of power overlooked?
Many labs send out a questionnaire as part of the quote process, but these often are ignored or result in limited answers unless the issue is pressed. Product requirements change from initial concept, and the information to the lab doesn’t get updated.
Anyone working in a lab has seen the deer-in-the-headlights look when it is realized that some of the susceptibility testing could damage the UUT. To avoid this, the required test suite should be examined and appropriate protection incorporated into the design, such as transient protection and filtering.
Performing a modification to incorporate a series component at the test lab can be a formidable task even for the most skilled technician. And if successful with that solution, a board spin will be needed to incorporate the component for production.
Don’t forget about protecting and decoupling the support equipment. Most test configurations can’t separate the support from the UUT without an intentional separation design.
For example, noisy support computers couple noise into the enclosure via the cables and reradiate at levels that are over the limit in an ambient test setup. Time often is spent during the test process to bring those elements under control. Secondly, and not insignificantly, the support equipment can easily be damaged if not protected.
Conclusion
For testing products to MIL-STD-461F as well as all EMC standards, planning and communications are vital. Prepare for the product test with the same diligence and commitment you used in designing it. Casual familiarity is not acceptable. Know the characteristics of your product and the requirements of your test plan. Then, when the test lab asks if you are ready to test, your answer will be a resounding yes.
About the Author
Steven G. Ferguson is vice president of operations at Washington Laboratories. He has been working in the compliance test arena for more than 35 years at test laboratories and manufacturing companies designing products, developing procedures, and performing tests. Mr. Ferguson also presents a hands-on course in testing to MIL-STD-461 for multiple government and industrial clients. Washington Laboratories, 7560 Lindbergh Dr., Gaithersburg, MD 20879, 301-216-1500, e-mail: [email protected]