Immunity Testing: Examining Requirements and Test Methods

Intertek Testing Services

The fact that man-made electromagnetic energy can also interfere with electric/electronic systems is a common experience. Anyone who has watched television at the same time that a vacuum cleaner connected to an adjacent plug is being used or switched off has realized that there is some kind of electromagnetic connection between the two.

Basic Concepts

Immunity is the ability of devices, equipment, or systems to perform without degradation in the presence of an electromagnetic disturbance. Susceptibility is lack of immunity (inability to perform without degradation).

In many cases, susceptibility is thought of as a temporary reduction of performance or degradation of quality. Even so, susceptible electronics may directly or indirectly concern safety. For example, a telephone call might be disturbed in such a way that the quality of the speech signal is affected, but it may also imply a blocking, and in the case of an emergency call, this may have disastrous consequences. Today we can easily find numerous examples of more or less serious electromagnetic problems:

Some tests may seem simple to perform, like using an ESD pistol to test for immunity against electrostatic discharges. It should be noted, however, that the discharge pulse represents a very fast transient, which is easily affected by cabling, ground planes, how the gun is held by the operator, etc. Full compliance testing therefore, unfortunately, tends to be a bit expensive.

That is why precompliance testing is recommended in order to find the most susceptible parts of electronic devices at an early stage of the development process. Simplified standard tests may be used for such pretesting. For example, if a computer-controlled apparatus does not cope with an electrostatic discharge toward an adjacent ground plane, the apparatus is probably not tight enough to pass a test with radiated electromagnetic fields. Immunity testing by injecting a fast transient on a cable can also be performed as a fairly simple precompliance test with any suitable pulse generator (for example, an ESD pistol instead of a burst generator). And if you don't have an ESD pistol, you might use your imagination—why not stroke an Angora cat!

Immunity test methods are described in detail in basic standards. My intention is therefore not to rewrite the standards in this article, but to highlight practical circumstances that I believe are worth noting. Furthermore, the focus will be on the commonly used immunity test methods according to the basic standards (Table II), to which generic and product-oriented standards refer.

EUT and Test Performance

When planning an immunity test, it is important to define the equipment under test (EUT), and its ports (see Figure 1). Input and output ports may be process measurement and control ports as well as certain data buses involved in control, such as a field bus. Signal ports may be signal lines and data buses. Power ports may be ac or dc. The enclosure port is defined by chassis and screening cables. Immunity testing means that typical electromagnetic disturbances are applied to all relevant ports while the apparatus is in operation.

Figure 1. The EUT may have many ports that might pick up disturbances.

The configuration, as well as operational modes of the EUT, should be carefully described in a test plan. The installation and the operational test conditions should be in accordance with product specifications. Certain tests have to be performed with the EUT placed on a reference ground plane. Test setups, with necessary insulating supports and ground connections, are usually well described in standards and should be followed in detail to obtain reproducible results.

Immunity testing is to be conducted in a well-planned and reproducible manner, as specified in a test plan. If the test object, in its normal operation, is to be connected to auxiliary devices, it should be tested with a minimum representative configuration of such devices. The most susceptible modes of operation should be selected and the configuration varied to achieve maximum susceptibility. Severity levels as well as criteria have to be defined. It is, of course, important to define the test setup in the test plan as well as to analyze what parameters have to be monitored during testing. Television cameras are often used to guard screens and control lamps and to monitor movements. Electrical measuring instruments, like oscilloscopes and bit analyzers, must be arranged in such a way that they cannot affect the test results. Television cameras must be shielded from radiated electromagnetic fields. When illuminating the EUT with electromagnetic fields, fiber-optic transmission techniques are often used for transferring the monitored signals into the control room. This reduces the possibility that the field is coupled onto the cabling.

The required severity test level/class for an apparatus or system depends on the environment and installation conditions where it will be used. Equipment to be used in a well-protected environment (Levels 1 and 2), like a computer room, may be tested at a lower immunity level than equipment to be used in an unprotected environment (Levels 3 and 4). Severity levels are defined in each basic standard. The following are generally used for testing against radiated RF fields (IEC 61000-4-3) and conducted disturbances induced by radiated fields (IEC 61000-4-6), respectively:

However, these figures should not be viewed as though an immunity test with an RF field of 3V/m will give identical results as an induced test of 3V if performed on the same object at the same frequency (see Figure 2). Different test methods are not directly convertible and should not be regarded as kind of an absolute measure of immunity. Standardized test methods may not simulate a real situation at all, but as previously mentioned, the test methods are well defined to achieve the highest possible reproducibility.

Figure 2. An example of comparative testing performed on the same test object with two different immunity test methods, RF radiation and injection respectively (IEC 61000-4-3 and 61000-4-6). The marked frequency intervals indicate interference.

The induced RF method may be used from the kHz region up to 230 MHz, while the RF field illumination method is used from 26 or 27 MHz up to GHz. In recent generic and product standards, testing with radiated fields is commonly used above 80 MHz, while the induced RF test method is used below 80 MHz. The reason is simple: the induced RF test method is much cheaper. Testing of immunity against radiated RF fields at 30 MHz (wavelength l = 300/f MHz = 10 m) requires a fairly large (length ~ 20 m) and expensive anechoic chamber. The wavelength of the radiated field as compared to the mechanical dimensions of the EUT must be taken into consideration with respect to the size of cabinets as well as the lengths of connected wires.

Test Conditions

The uncertainty of the actual value of the ultimate disturbance level creates the need for margins. All instruments in the chain needed to define a specified level of disturbance must be calibrated (for example, oscilloscopes for measuring pulse characteristics and probes and/or power meters used for defining field strength). If, for example, the total uncertainty when reproducing an RF field is about 30% (including monitoring), the field strength control should be set on 13 V/m if 10 V/m is to be given as the minimum test level.

Most basic immunity standards define tolerances for disturbance characteristics like pulse and rise times, and pulse width. Besides calibration of instruments used for reproducing disturbances, regular checks are recommended to find out (at an early stage) if a generator or a measuring instrument tends to go out of specifications. Transient pulses can be monitored with an oscilloscope. Note, however, that to study a pulse with a rise time shorter than 1 ns, you need an oscilloscope having a real-time bandwidth of at least one GHz, since a rise time t_r ~0.7 ns corresponds to a 3-dB bandwidth B ~500 MHz (B = 1/ t_r).

Environmental conditions may affect some of the immunity tests. The physical properties of an electrostatic discharge, for example, depend to a great extent on humidity. The performance of transient generators may also, to some extent, depend on humidity, particularly for an old burst generator using a spark gap. Requirements concerning environmental conditions therefore apply to some of the immunity standards. According to relevant standards, the air temperature usually has to be controlled to within 15°–35°C, air pressure within 86–106 kPa, and humidity according to the following:

Severity Criteria

Test results are classified as pass or no-pass on the basis of the operating conditions and functional specifications of the EUT, according to the following general criteria:

Detailed specifications are left open to the manufacturer and the parties involved in testing, who therefore must analyze relevant parameters carefully before testing. It is not, for example, a major concern if a PC screen flickers during a static discharge, as long as the PC is running normally after the disturbance has ceased. The needle of a sewing machine should not, on the other hand, be allowed to move due to a discharge. A measuring instrument or a gauge might go out of specs during a disturbance with short duration, but should stay within its specified error band when a continuous disturbance source, like a radio transmitter, is used in its vicinity. A household machine with an automatic power control should not be affected by a mobile phone in such a way that it may cause harm, but a power variation of 10% is probably harmless.

Degradation or loss of function, which is not recoverable due to damage of equipment (components and/or software), or loss of data, should never be accepted.

The collateral standard for medical devices, IEC 601-1-2 (EN 60601-1-2), states that the EUT shall perform its intended function as specified by the manufacturer or fail without creating a safety hazard. One could say, therefore, that a medical device is allowed to break during a test, provided that the alarm system is still in operation! This, of course, is not a recommended practice, and this paragraph will be replaced by detailed criteria in the forthcoming revised collateral standard.

Product-related standards usually specify operating conditions and criteria for specific products. In some cases, such criteria might seem to concern product quality rather than safety. For example, the immunity performance criteria for a television according to CISPR 20 is merely a perceptible degradation of the picture.

No deviations are allowed when harmonized standards are used for compliance, but in certain cases the test requirements may be reduced! As an example, the generic immunity standards state that: "It may be determined from consideration of the electrical characteristics and usage of a particular apparatus that some of the tests are inappropriate and therefore unnecessary. In such a case it is required that the decision not to test be recorded in the test report."

Note, on the other hand, that the generic standard may also require enhanced severity levels: "In special cases situations will arise where the level of disturbance may exceed the levels specified in this standard, e.g. where a handheld transmitter is used in proximity to an apparatus. In this instance special mitigation measures may have to be employed."

Standard Tests and Their Mirrored Reality

Standard immunity tests concern continuous or transient disturbances, low- or high-frequency phenomena, and conducted and radiated electromagnetic effects. The goal of testing is to reveal eventual susceptibility against common disturbances, but observe that in some cases the equipment may be damaged. This might be the case when it comes to testing of immunity against surges and electrostatic discharges.

Testing of immunity against radiated disturbances is usually performed by illuminating the EUT with an electromagnetic field. Interference might also be simulated by directly inducing disturbances on cables connected to the equipment.

Conducted disturbances are usually injected directly on an EUT port or injected/induced through an impedance, a clamp, or a coupling network (decoupling networks are not used to influence auxiliary equipment). Generic, general, and product family standards may stipulate that certain tests are unnecessary if the cables connected to the EUT are short; therefore, injected RF disturbances and fast transients are required only if the total length according to the manufacturer's functional specification may exceed 3 m.

Continuous disturbances have to be applied as long as it takes for the EUT to fulfill a certain operation cycle. Concerning transient tests, interference usually depends on when the pulse occurs in respect to the operational mode or a clock cycle. It is therefore necessary to repeat the test a number of times in order to catch the worst case.

Testing with Continuous Disturbances

Testing with radiated electromagnetic fields according to IEC 61000-4-3 simulates possible radio communications threats. Testing is usually done within the frequency range up to 1 GHz with a field strength of 3 V/m (Level 2) or 10 V/m (Level 3). However, if one wants to be sure that an apparatus will not be affected by, for example, a mobile phone in close vicinity, a test with fields of at least 30 V/m is recommended (100 V/m for life-supporting equipment and critical apparatus).

Modern standards, like the European generic standard EN 50082-1: 1992, may still refer to old "basic" standards like IEC 801-3, which actually restricts testing up to a maximum of 500 MHz with an unmodulated carrier. It should be noted that 500 MHz is far too low to cover today's mobile threats. The generic standard, however, has added a requirement for amplitude modulation of the electromagnetic field. This is usually far more severe than testing with an unmodulated carrier, since it is mainly the envelope of the picked-up electric signal that causes the interference in electronic circuits. Most modern standards therefore require AM and/or PM modulation (FM has little effect and is therefore not used).

The high-frequency carrier couples the electromagnetic wave into electronic equipment with the help of different "antennas," such as cables and enclosure parts of suitable length. For example, a cable of 1 m length can act as a monopole antenna with a maximum sensitivity around 75 MHz (l/4 = 1 m). After being picked up via antenna effects, a radio frequency signal can itself be a threat to an electronic circuit. For example, a selective HF receiver, as in a navigation system, is sensitive to RF frequencies within its passband. An RF signal may also cause resonances of a few hundred MHz, due to lead inductances and capacitors (_res = 1/2(check)LC). Nonlinear effects in the electronics, however, usually rectify or demodulate the RF signal. After the high-frequency carrier is eliminated, the remaining disturbance may represent an offset or a low-frequency signal, which might affect a sensitive analog amplifier or an integrating A/D convertor. The reason a mobile phone of type TDMA can affect audio equipment is because the carrier is keyed by an LF signal (217 Hz in the case of GSM), which together with harmonics causes disturbances in the acoustic range (up to some kHz).

Besides modulation, IEC 61000-4-3 also requires that testing be performed with a homogenous EM field. This implies high-performance test facilities, with a need for a high-power amplifier, since the antenna has to be placed at a relevant distance from the test object.

Accurate RF immunity testing requires a shielded anechoic chamber, but can also be performed in a large GTEM cell. For testing according to CISPR 20/EN 55020, an open stripline is prescribed, a simpler form of a TEM cell that is used for generating fields within the frequency range 150 kHz–150 MHz. Because this is a rather specific test method for immunity testing of radio, television, and audio equipment, it will not be addressed here.

Testing with induced RF disturbances simulates how electromagnetic fields may be picked up by cables, which in turn feed the disturbances into the connected equipment. Induced RF immunity testing according to IEC 61000-4-6 has been introduced as a less expensive alternative to RF field immunity testing in the lower frequency range. Standards primarily refer to testing with a modulated RF carrier from 150 kHz up to 80 MHz, in some cases up to 230 MHz. The current into the EUT is injected through a coupling network or by a clamp.

The major advantage of this method compared to RF field testing is that there are no requirements for using an anechoic chamber. Testing is usually performed by injecting one cable at a time. This technique is subject to criticism, however, because in a real situation when there are several cables connected to the EUT, the whole system will be affected by an electromagnetic field and all cables may carry RF currents simultaneously. Immunity testing requires that all ports be subjected to different threats, such as injection of transient disturbances on ports and cables, or illuminating chassis ports and cables with electromagnetic fields.

Testing with LF magnetic fields is defined in the basic standard IEC 61000-4-8. The test simulates magnetic fields resulting from currents running in power-line systems (there are also two basic standards for testing of immunity against pulsed magnetic fields, IEC 61000-4-9 and -10).

Earlier generic and product family standards do not refer to this fairly new basic standard, but may have general requirements concerning immunity against power-line–frequency magnetic fields of from 3 A/m up to 300 A/m. However, testing is only required when it comes to equipment, including a component or a subunit, which is known to be sensitive to LF magnetic fields, for example, a Hall element, a coil, or a monitor. Immunity testing against LF magnetic fields up to around 100 kHz is also stipulated in some other standards, like MIL-STD 462, RS-101 (30 Hz–100 kHz).

Immunity Against Transients and Other Disturbances of Short Duration

ESD tests are performed with an ESD pistol. This is a pulse generator with a capacitor that can be charged to about 15 kV and which is discharged via a known impedance. The test level is often 4–6 kV, contact, and/or 8 kV, air discharge. However, discharges up to 10–20 kV may occur in real life, especially during the winter in northern parts of the world when the air humidity drops to very low levels due to indoor heating; this can also occur in other dry parts of the world like desert areas. Note that the humidity may also be very low close to a surface that is heated by electronic components.

Note that there are two ESD standards, IEC 801-2: 1984 and IEC 801-2: 1991 (equivalent to IEC 61000-4-2). The first, edition 1, defines the discharge pulse with rise time to 5 ns, while the rise time according to the latter, edition 2, is only 0.7 ns, which is far quicker and usually represents a more severe threat. It should be noted, however, that both pulses exist in reality. The slower air discharge is typical for a finger discharge against an object (human body model), while the quicker pulse is typical for a discharge through a sharp conductive object, like a screwdriver touching a metallic enclosure (charged device model) (see Figures 3 and 4).

Figure 3. An electrostatic discharge in real life may have a rise time of about 5 ns, as defined in the old ³basic² standard IEC 801-2: 1984, modeled on a discharge through a human finger.

Figure 4. If the discharge is caused by a charged metallic object like a screwdriver, the rise time may be as quick as 0.7 ns, as defined by the new standard IEC 801-2:1991 (IEC 61000-4-2).

It is generally the case that the higher the frequency, the shorter the rise time of a disturbance, and therefore the easier it is to couple into an electronic circuit. It happens, however, that equipment can be immune to the shorter pulse but not to the slower pulse. This can be explained by the fact that faster pulses are more easily bypassed by stray capacitances.

Even though a short ESD pulse may have low energy, it may still damage sensitive integrated circuits (ICs). Device failure may occur some time after the original damage was sustained; for example, a pn-channel in an IC may be affected in such a way that it will be damaged after being in operation for some time. Because the IC may leave the factory in this undiscovered damaged state, it is said to carry a latent defect, known to industry as "the walking wounded."

Burst or electromagnetic fast transients (EFT) represent transients generated from circuit breakers, nonprotected relays, etc. IEC 61000-4-4 defines an idealized burst with relatively long duration (15 ms) repeated at a rate of a few kHz. The pulse amplitude of the transients during the burst is usually 0.5–4 kV peak, but since the duration of each transient is fairly short, the bursts have fairly low energy. The pulse rise time is typically 5 ns and the pulse width, 50 ns. They are generated with a burst generator and injected into cables through a coupling network or a capacitive clamp.

The bursts are injected into both power lines (both common-mode [line to ground] and differential-mode [line to line]) and I/O lines (if longer than 3 meters). Power-line injection is performed using a discrete capacitor, while I/O injection is performed using a long capacitive clamp (with the capacitance distributed along the cable).

Surge is a transient voltage propagating along a line. Surges can be induced in cables by lightning, but might also pop up on the power line when connecting a phase compensating capacitor into the distribution net or when a fuse breaks. The double exponential pulse, according to IEC 61000-4-5, is defined by both open-circuit and short-circuit characteristics. It is usually generated with a hybrid generator, capable of giving a peak voltage of 1–4 kV and a maximum peak current of some kA. The surge pulse has a fairly long pulse-width, 20 or 50 µs, and therefore represents a fairly high energy pulse that may easily damage unprotected electronic circuits and components.

Surges are to be applied on power lines with a defined timing referred to the power-line phase. They may also be imposed on some signal cables, like long telecom cables. The pulse is injected through an impedance or a network.

Voltage variations, dips, and short interruptions are caused by faults in the distribution system or by a sudden large change of load. IEC standard 61000-4-11 defines simulation of typical power tripouts with a duration of a few cycles of the power-line frequency.