EMC Problems in Switching Power Supplies and How to Solve Them
How switching power supplies, the biggest source of EMC problems in electronic products, generate noise and what to do about it in design.
What an EMC test plan should contain before you go to the certification lab: standards, operating modes and performance criteria.
Many products going to EMC testing lose time not because of a technical problem but because of incomplete preparation. On test day there is debate about which standard applies, the product's operating mode is unclear, the functions to monitor have not been defined, or the necessary auxiliary equipment has not been brought. These uncertainties increase the cost of every hour spent in the laboratory.
A well-prepared test plan prevents all of these problems. In this article we explain why a test plan is needed, what information it should contain, how to determine the standards to apply, how to define operating modes, the test setup and performance criteria and how to coordinate with the laboratory.
A test plan is the document defining which tests the product will undergo, at which levels and under which conditions, and how results will be evaluated. This document creates a shared understanding between manufacturer and test laboratory and removes in advance differences of interpretation that could arise during the test.
A test plan also ensures repeatability. When the product is retested later, for example after a design change, results can be compared directly using the same plan. The test plan is also an important part of the declaration of conformity and technical file; in audits it shows how the product was tested.
The first step in the test plan is determining the standards that apply to the product. This is done according to the product type, the environment of use and the markets where it will be sold. If there is a specific product standard, for example CISPR 32 and CISPR 35 for multimedia equipment, that standard applies first. If there is no specific standard, generic standards are used according to the environment.
Once the standards are determined, the tests each standard applies to the product and at which levels are listed. The tests to be applied to each port the product has, power input, signal and communication ports and the enclosure, are specified separately. If more than one market is targeted, a test list meeting the requirements of the different markets together is created.
The product should be described in detail in the test plan: model name, hardware and software version, supply voltage and frequency, dimensions and weight. If the product has different variants, it should be explained which variant will be tested and how it represents the others. This information also appears in the test report.
The auxiliary equipment the product will run with during the test should also be defined. Auxiliary equipment such as computers, load devices, sensor simulators and communication partners should be chosen to represent the product's real use. Whether this equipment will be inside or outside the test area and how it will be monitored during the test should also be stated.
The mode in which the product runs during the test directly affects the results. In emission tests the product should run in the mode in which it produces the most noise; in immunity tests in the mode in which it is most sensitive. These modes are not always the same, and some products may need more than one mode tested. The reasoning behind the mode choice should be explained in the plan.
The test setup, the arrangement of the product, cables and auxiliary equipment in the test area, is planned according to the rules defined in the standards. Different setups are used for tabletop and floor-standing products. Cable length, arrangement and termination significantly affect the result, especially in radiated emission tests. The setup should therefore be shown in the plan with a drawing or photograph.
In immunity tests, whether the product passes is evaluated against performance criteria. Standards usually define three criteria, A, B and C; but what these criteria mean concretely for the product must be defined by the manufacturer. For example, criterion A for a measuring instrument could be defined as the measurement error not exceeding a certain percentage.
The plan should clearly state which functions will be monitored during the test, how they will be monitored and what acceptable behaviour is. Is a brief flicker of the display, a lost packet in communication or an LED blinking acceptable? These questions should be answered beforehand, not on test day. Otherwise every behaviour seen during the test becomes a subject of debate.
Once prepared, the test plan should be shared with the laboratory and its views obtained. The laboratory can point out incomplete or unclear points in the plan, estimate test time and cost more accurately and prepare the necessary equipment in advance. This coordination makes test day run efficiently.
It is recommended to bring to test day a spare sample, the necessary auxiliary equipment, cables, software tools and, for quick intervention if a problem arises, filter components and ferrites. Having an engineer who knows the product well present during the test also means problems can be understood and solved quickly.
Preparing a simplified test plan for in-house pre-tests is also very useful. The aim of pre-testing is to find in advance the problems that will be met in the certification test; so pre-testing should be done under conditions as similar as possible to the certification test. The same plan can serve as the basis for both pre-testing and certification testing.
Pre-test results are also a good way to reveal uncertainties in the plan. For example, which mode is noisiest or which function is most sensitive becomes clear in pre-testing, and the plan is updated accordingly. This work can be done quickly in a pre-compliance chamber. For test infrastructure options, see our semi-anechoic chamber page.
The laboratory can help, but the manufacturer knows the product's operating modes and acceptable behaviour best. The plan works best when manufacturer and laboratory work on it together.
Yes, but changes should be documented and justified. Changes made during the test in particular should be clearly stated in the test report.
It is recommended. A simplified plan makes pre-test results comparable with the certification test.
The A, B and C criteria of the standards are turned into concrete, measurable statements for the product's monitored functions and set in writing before the test.
The mode producing the most noise is chosen for emission tests, and the most sensitive mode for immunity tests. More than one mode is tested if necessary.
A spare sample, auxiliary equipment, cables, software tools and filter components and ferrites for quick fixes; an engineer who knows the product well should be present.
Your free site survey and quote will be ready within 24 hours.
How switching power supplies, the biggest source of EMC problems in electronic products, generate noise and what to do about it in design.
Rules for ground planes, layout, filtering and connectors that prevent EMC problems at the source when designing printed circuit boards.
How the noise a product radiates is measured in a semi-anechoic chamber, with the pre-scan, maximization and final measurement steps.