EMC Test Items Explained: What Each Test Applies to a Product
Every product that is sold into a regulated market is measured against a list of disturbances, and the list is fixed well before the design starts. Knowing what each item applies to the product, and how it stresses the board, turns compliance from a test-day surprise into a set of requirements that the layout can be checked against. The EMC test list divides into two families: what the product emits, and what the product tolerates.
The Emission Family
Emission tests measure what the product sends out into the world, and there are four items in common use.
The radiated emission test measures the energy that leaves the product through space, using an antenna in an anechoic or semi-anechoic chamber at a defined distance, usually three metres. This is the test that most engineers expect, and it is the one whose result depends most heavily on the layout of the board inside the product.
The conducted emission test measures the energy that leaves along the cables: the power cord, the control lines and the signal lines. The measurement is made with a probe or with a stabilising network that separates the disturbance on the cable from the mains supply. Because the cable is the longest conductor in most products, this test often fails when the radiated test passes, and the fix is usually in the grounding and filtering at the interface rather than in the layout.
The harmonic current test measures the distortion that the product injects back into the mains supply, as a series of harmonics of the line frequency. It is a property of the power supply rather than of the logic, and it is addressed by the choice of input stage and by power factor correction rather than by anything on the board.
The flicker measurement records how much the line voltage changes as the product draws its current, and whether that variation would produce a visible flicker in lighting on the same supply. Like the harmonic test, it is dominated by the power stage.
The Immunity Family
Immunity tests apply a disturbance and check that the product continues to work. The list is longer, and each item stresses something different.
An ESD immunity test discharges a charged source into the product, either by direct contact with a conductive surface or through the air to an insulated surface. The discharge rises in less than a nanosecond and carries a current of several amperes, so it is a high-frequency event with very little total energy. On a board it stresses the connector shells, the enclosure seams, the mounting points and any trace or plane that runs close to the edge. The current path it takes through the product is what matters, and the design intent is to give it a path to the chassis that does not pass through a circuit.
The radiated immunity test applies a radio frequency field to the product and checks that it keeps working. The frequency range and the field strength are defined by the applicable standard, and the failure mode is usually demodulation: a fast edge on a cable, induced by the field, arrives at a device input and is interpreted as data.
The conducted immunity test applies radio frequency energy to the cables rather than through the air. This is the test that finds filters that are ineffective because their ground reference is far away, and it is the reason filters should be placed at the connector rather than at the device.
The voltage dip and interruption test reduces the supply voltage, or removes it briefly, or varies it continuously. The product is expected to survive a short interruption without losing stored data and to recover when the supply returns. Devastating on products with a processor and a memory that is written continuously, it is addressed in firmware as much as in hardware.
The surge test applies a high energy pulse that models a lightning strike or a switching event on the grid. The standard waveform has an open circuit voltage rising over about 1.2 microseconds and falling to half in 50 microseconds, with a short circuit current rising over 8 microseconds and falling in 20. The characteristics that follow from those numbers are a slow rise, a long duration and a large amount of energy, so the protection has to absorb energy rather than merely respond quickly.
The electrical fast transient test applies a burst of very fast pulses instead: a rise time of about 5 nanoseconds, a pulse width of 50 nanoseconds, a burst lasting around 15 milliseconds, and a burst period of about 300 milliseconds. The characteristics are the opposite of the surge test: a fast rise, a short duration, a small amount of energy, and a high repetition rate. The failure mechanism is different as well, because the disturbance couples capacitively through insulation rather than arriving as a conducted current, so the fix is often a ground plane and a filter rather than a bigger protection device.
A power frequency magnetic field test is applied where the product may operate near transformers or heavy conductors. It applies a steady field at the line frequency, and the design response is to reduce the area of any loop that could pick it up.
How the Tests Are Run
The environment matters as much as the item. Radiated work is done in a chamber, which removes reflections so that the measured field is the product´s own. Conducted work is done with a stabilising network that defines the impedance the product sees on the mains, so the measurement is repeatable between laboratories. The ESD gun, the surge generator and the burst generator each have their own coupling networks, and the way the product is earthed and cabled during the test is specified by the standard because it changes the result.
The configuration is part of the test. A product is tested as it will be installed, with its cables dressed as they will be routed and its covers fitted as they will be fitted. Testing a bare board on a bench gives a result that has no relation to the compliance measurement, and testing with a different cable arrangement in each session produces results that cannot be compared with each other.
Preparing the Design
The useful preparation is to map each test item to the feature on the board that determines the result.
For radiated emission, the feature is the loop area of the fast nets and the continuity of their return paths, plus the placement of the clock and the switching supply relative to the cables. For conducted emission, the feature is the filter and the ground reference at the connector. For ESD, the feature is the chassis path and the clearance around the board edge. For radiated and conducted immunity, it is the loop areas and the filters again, seen from the other direction. For surge, it is the energy rating of the protection device and the width of the traces that carry the pulse. For fast transients, it is the plane structure and the speed of the protection device.
That mapping is worth writing down before the layout, because it converts a test that happens once into a list of checks that can be applied while the design is still open. Our design review uses that approach, and for products that are certified after assembly, the configuration that was tested is held with the build record so that a production unit matches it; the traceability for that is part of our quality management system.
Where a product is developed with a partner, the test plan is worth agreeing at the same time as the layout. Pre-compliance work on a prototype is cheaper than a failed formal test, and functional testing at the prototype stage helps confirm that the board behaves as expected before the chamber is booked, which is where our PCBA testing service supports the development process.
FAQ
Which test fails most often? Conducted emission on the mains cable and radiated emission at the clock harmonics are the two most common failures in small products.
Is the surge test the same as the burst test? No. Surge is a single high energy pulse with a slow rise; the fast transient burst is a train of low energy pulses with a fast rise and a high repetition rate, and they require different protection.
Does the harmonic test depend on the board design? Rarely. It is determined by the input stage of the power supply and by the load profile.
