In situ EMC testing--Immunity

Practical aspects of immunity tests

In contrast to emissions tests, the basic feature of an immunity test is that the EUT is subjected to external interference phenomena generated by the tester, and its performance is monitored for unusual or unacceptable operation. But, as in the case of emissions, not all the standard methods can be properly applied in-situ. So the issues that have to be addressed for in-situ immunity testing are:

++ is it practical and legal to generate the necessary disturbance levels, and do special precautions need to be taken?

++ will the disturbances generated by the tests interfere with or upset other equipment than the EUT?

++ can the standard measurement method be applied successfully and correctly; if not, what are the implications of necessary departures from the standard method?

The latter question is identical to that already raised by emissions tests, and has been commented on. The question of precautions against interfering with other apparatus, either within the installation or belonging to an innocent third party, is in some respects the reciprocal of the problem of ambients. Since it presents different problems for different phenomena, it is treated separately for each, below.

Electrostatic discharge

--- The ESD event

The ESD phenomenon

Electrostatic discharge is a source of transient upset which most typically occurs when a person or other body that has been charged to a high potential by movement across an insulating surface then touches an ground/earthed piece of equipment, thereby discharging through the equipment. Currents of tens of amps can flow for a short period with a very fast (sub-nanosecond) risetime. Even though it may have low energy and be conducted to ground through the equipment case, such a current pulse couples very easily into the internal circuitry and can disrupt its operation.

When two objects with a high electrostatic potential difference approach, the air gap between them breaks down and the charge difference is equalized by the current flow across the resultant ionized path. The current route is completed by stray capacitance between the objects and their surroundings, and the inductance and resistance of the path. The actual current waveform is complex and depends on many variables. These include speed and angle of approach, and environmental conditions, as well as the effect of the distributed circuit reactances. At lower voltages a "precursor" spike due to the local area of the source (a finger or a metal tool) is produced which has a very fast risetime, of the order of a few hundred picoseconds. This spike, although low in energy, can be more damaging to the operation of fast digital equipment than the bulk discharge that follows it, which may have a risetime of 5-15 nanoseconds.

The discharge current will take the route of least inductance. If the case is well bonded to ground then this will be the natural sink point. If it is not, or if it is non- conductive, then other routes may be via the connecting cables. Because the discharge edge has an extremely fast risetime (sub-nanosecond), stray capacitive coupling is essentially transparent to it, whilst even short ground connectors of a few nH will present a high impedance. Apertures in a conductive enclosure will also result in intense local magnetic fields which will couple to the internal circuits.

The test method

--- ESD test methods

The standard test for ESD immunity is defined in IEC 61000-4-2. This requires a well-defined test set-up with a ground reference plane and horizontal and vertical coupling planes; the EUT is placed a specified distance from the reference plane and discharges are applied to it from a calibrated generator, at specified test levels. The generator voltage is developed with respect to the reference plane.

Some product or generic standards call up different versions of the basic ESD test standard, and this guide describes the issue (1995) which was current at the time of writing. It is generally agreed that this issue is more realistic, comprehensive, and tougher than previous issues of this basic standard. A successful test to the 1995 issue thus gives confidence that a successful result would have been achieved if an earlier issue had been employed, but this cannot be guaranteed.

The standard defines two methods of application, known as contact discharge and air discharge. In the air discharge method, the point of the discharge generator is brought close to the EUT until an arc occurs between the two.

The human body is modeled as a 150pF capacitor charged to the test voltage, and discharged into the EUT through a 150 -ohm resistor and a discharge tip which simulates the dimensions of an outstretched finger. The return path is completed by a strap from the test capacitor to the ground plane. For maximum repeatability the tip should approach the EUT as fast as possible and with the probe at right angles to the surface.

To improve repeatability the contact discharge method requires a test probe in direct contact with the EUT or with a nearby coupling plane, and the test discharge current is initiated using a high-voltage vacuum relay. This avoids the unpredictability associated with variable approach speed and environmental conditions. The tip itself is sharply pointed to allow a clean contact with the EUT surface by penetrating any surface corrosion or coatings which are not intended to have insulating properties.

Coatings which are declared to be insulating are subjected to the air discharge only.

Note that the air and contact discharge methods are not alternatives but complementary. A correlation between the results of the two methods has not been possible, and it cannot be assumed that their test severities are equivalent, although this is the implication of their widespread use in product standards.

In-situ tests

IEC 61000-4-2 gives some limited guidance for in-situ ESD tests, which it calls "post- installation" tests. As with the conducted emissions test, the most important aspect is the provision of an adequate ground reference connection. The standard requires a specific plane placed on the floor of the installation about 10cm from the EUT. It should be of copper or aluminum not less than 0.25mm thick, with dimensions 0.3m wide and 2m length where possible. It is connected to the protective ground/earthing system, or to the ground/earth terminal of the EUT, if this exists. The discharge return cable of the ESD generator is then connected to this reference plane at a point close to the EUT (a typical set-up).

--- Typical in-situ ESD test set-up (after IEC 61000-4-2)

Choice of application points

Contact discharges should be applied to metallic points accessible to the operator or user during normal use. This excludes servicing operations (such as changing pcbs) that would be carried out by trained maintenance personnel who can be expected to take proper anti-static precautions. A draft amendment has been proposed which offers a definition of accessible parts. In brief, excluded from testing would be:

++ points that are only accessible under maintenance

++ points that are accessible under customer's service

++ points and surfaces that are no longer accessible after installation

++ contact pins of connectors that have a metallic connector shell - contacts within a plastic connector should be tested by air discharge only

++ contact pins which are ESD sensitive because of functional reasons and that have an ESD warning label

++ accessible parts that are only accessible when the equipment is not fully installed.

Air discharges are attempted at slots, seams, insulated switch and keypad apertures and other areas where a discharge would be expected to occur but which cannot be adequately tested by the contact method. The discharge voltage is maintained on the round tip as it is approached at right angles to the EUT as fast as possible without causing damage. The actual points of application must be chosen experimentally by investigating all around the boundary of the EUT with both direct and indirect discharges. If the EUT is satisfactorily immune at the scheduled test levels, these should be increased to attempt to find the worst case application points.

Testing at lower and higher levels

It is well known that on this test, upsets can occur at some stress levels but not occur at higher stress levels. Testing solely at the maximum stress level specified by the product or generic standard may thus provide a misleading result. A fraction of the maximum test level should be applied at first, rising in increasing fractions until the maximum is achieved.

It may also be desired to apply an even higher test level to determine the "margin" that has been achieved, but this should only be undertaken where the possibility of damage to the equipment under test and consequent financial or time losses have been fully evaluated and found acceptable.

Quick testing method to identify worst-case points

Most ESD test "guns" have a rapid-fire option for air discharge, in which their output relay is disabled and their internal high-voltage generator connected directly to their probe tip. They will then discharge as soon as their tip charges sufficiently to break down the air gap to the equipment under test, resulting in a rapid succession of discharges. Varying the distance of the probe from the tested point varies the breakdown voltage, allowing a whole series of tests over the full range of voltages to be carried out in a few seconds. This method may be used to quickly identify the most susceptible test points and voltages, so that they may be more properly tested according to the standard, and is especially useful where there are a large number of such points.

The maximum distance of the tip from the tested point should not exceed that which is seen to cause air breakdown at the maximum test voltage, unless some overtesting is required. Because the test method is not carefully controlled, it is likely that some discharges will exceed the maximum level, so care should be taken (which might include not using this time-saving method) where damage to the tested unit is possible and must be avoided.

Effect on other equipment

The ESD pulse has low energy but very wide bandwidth, and it couples via the loop formed by the ESD generator, the EUT, the ground reference plane and the generator's return strap. Therefore it radiates very effectively. Equipment in the vicinity, especially digital processing equipment with poor immunity, may be upset by the pulse. Effects are possible to a distance of 10m but unlikely beyond that, unless a very sensitive system is involved. Clicks will also be heard on radio reception over a wide radius, but as they are single pulses with a low repetition rate they are unlikely to be intrusive.

If possible, the tests should be done in an area removed from other operating equipment, or if this is not possible, nearby equipment should be powered down. In the worst case where neither of these are possible, operators of adjacent equipment should be warned to expect unreliable operation for the duration of the test.

Electrical fast transient bursts

The EFT phenomenon

When a current carrying circuit is interrupted by an unsuppressed air gap switch (such as a contactor or relay) the resultant voltage rise across the gap causes a repetitive ignition and extinction of an arc discharge - the so-called "showering arc" effect. The nature of the arc is determined by the magnitude of interrupted current, the circuit and load inductance and stray capacitance, and the rate of separation of the contacts. It generally results in a burst of fast, low energy current transients which couple along the supply circuit or are radiated from the conductors on either side of the switch. The burst repetition rate varies from 10kHz to 1MHz.

The duration and rise time of these transients is short compared to the travel time in building wiring systems. Typically the rise time of each strike is less than 10ns and the duration is tens to hundreds of ns; transit time is roughly 5ns/m. Thus the supply wiring appears as a transmission line, so that the line characteristic impedance determines the transient source impedance, and the transients themselves are easily attenuated by distance. As a result, these fast switching transients have low energy and are only a threat to local susceptible equipment. But since there are many sources of such transients in the typical electrical installation, and because fast-rise events have considerable upset potential for digital circuits (though little or no potential for damage), equipment reliability will be enhanced by taking them into account.

While switching operations are the most widespread source, certain other similar sources- particularly arc welding and unsuppressed DC motors- can be particularly aggressive to equipment in their neighborhood.

Test method

The immunity test against EFT bursts is specified in IEC 61000-4-4. This calls for a generator with a particular pulse waveform, known as the 5/50ns waveshape because of its rise and fall times. The pulse is repeated in a burst at a repetition rate of 5kHz for a duration of 15ms, and this burst is repeated every 300ms for the test duration of one minute. Various lines have this burst signal applied, the combinations, polarities and levels being dictated by the applicable product or generic standard.

The test is intended to demonstrate the immunity of equipment when subjected to types of transient interference as described above. It is not intended to directly simulate such interference; this would be impossible because of the wide variety of interference frequencies, waveforms and coupling routes. Rather, the rationale for applying this test is that equipment which survives it has been found to be more robust in the presence of real sources of interference.

Coupling methods

Because the transients have a very high frequency spectral content, the test set-up and coupling methods must use RF techniques as with the electrostatic discharge set-up. A ground reference plane is mandatory and the test waveform generator and coupling network must be referred to it, as should the protective ground/earth. It should project beyond the EUT by at least 10cm on all sides. As with the ESD test, the EUT must only be ground/earthed according to its normal installation practice; free standing items should be spaced from the ground plane by an insulating support 10cm thick, while table-top equipment should be 0.8m above it. The basic principles of the laboratory EFT test are shown.

--- EFT test set-up in laboratory ground/earthed according to installation practice: coupling/"network~; L coupling clamp; signal ports; mains port; EUT fixed; distance

Power leads (less than 1 m long) are connected to the test pulse generator via a coupling/decoupling network which capacitively couples the transient burst onto each individual line. The coupling point is isolated from the supply by an L-C decoupling network in each line. This includes the protective ground/earth conductor, which has transients applied to it as well. Thus unless the EUT's enclosure is normally separately bonded to installation ground/earth, it too will see transient voltages.

There is some confusion amongst test laboratories as to the power conductors and their combinations to be tested, due to the different forms of words used in the generic/product standards and in IEC 61000-4-4. In some situations it may be that one power conductor or combination of conductors is less or more susceptible to the test than others, and this can lead to complaints of over-testing or under-testing. To be inclusive, and possibly overtest, all the phase conductors should be tested individually (e.g. L, N, E), and then all the phase conductors should be tested together (L+N), and then all the phase conductors and the unit's ground/earth should be tested (L+N+E). I/O cables are subjected to transient coupling via a capacitive coupling clamp spaced 10cm above the ground plane, which sandwiches the cable in question between two flat metal plates, both of which carry the test pulse burst referred to the ground plane. The distributed capacitance of this arrangement is 50-200pF. Note that because the voltage is applied through a capacitive divider (coupling clamp capacitance or coupling network capacitance in series with the EUT's capacitance to ground) the actual voltage applied to the EUT will be less than the peak open circuit voltage, which is specified as the test level.

In-situ testing

"Post-installation" tests are covered in reasonable detail in IEC 61000-4-4. The most significant difference compared to the laboratory test is that the burst generator is coupled directly into the power supply terminals, and the protective ground/earth terminal, at the EUT. No coupling/decoupling network is used, though a 33nF blocking capacitor may be needed. A reference ground plane of similar construction to that used for the ESD test, but this time of dimensions lm x lm, is laid "near" the EUT and connected to safety ground/earth at the power supply mains outlet; the burst generator is located on, and bonded to, this plane, and connected to the EUT terminals by a lead of less than l m length.

I/O lines are tested with the capacitive clamp as for the standard method. If there is insufficient room to use the clamp, it can be replaced by a length of conductive tape or foil wrapped around the cable loom under test, or by discrete 100pF capacitors connected to the terminals. IEC 61000-4-4 accepts that the results from each of these methods might be different, and suggests that the test levels might be amended "in order to take significant installation characteristics into consideration". As well as the coupling method, further variability is likely to be introduced by the detail of the connection of the burst generator to ground/earth reference in the vicinity of the coupling point, since this ground/earth reference is undefined except by the installation layout.

Although the fast transient burst test is a useful and widely referenced method of determining immunity to an important phenomenon, the variations and site-specific issues inherent in an in-situ test make it unreliable as a means of declaring compliance.

Any testing that is done should be meticulously reported, including a detailed and careful justification for the layout and choice of coupling method employed.

Interference with other equipment

The lack of any form of decoupling of the injected pulses, either onto the power terminals or onto I/O cables, makes this test quite aggressive with respect to other apparatus that is not being tested. Any equipment which shares the same power supply will see the same pulses, attenuated only by their distance from the point of injection.

The same is true of ancillary apparatus that is connected to any I/O cables under test.

Any other cables that run near to the cables under test must be considered as themselves being subject to the test, since crosstalk coupling will be effective for these wideband interference pulses.

Computers and programmable electronic systems such as PLCs and "intelligent" controllers, which have poor immunity can suffer upsets from EFT tests, with possible consequences as outlined above for ESD testing. Thus, as with the ESD test, nearby equipment and that which is connected to the tested circuit should be powered down. If this is impossible, operators of adjacent equipment should be warned to expect unreliable operation for the duration of the test.

Surges

High energy transients are generally a result of lightning coupling to the supply network, or are due to major power system disturbances such as fault clearing or capacitor bank switching. Lightning can produce very powerful surges by the following mechanisms :

++ direct strike to primary or secondary circuits: the latter can be expected to destroy protective devices and connected equipment; the former will pass through the service transformers either by capacitive or transformer coupling or insulation breakdown;

++ indirect cloud-to- ground/earth or cloud-to-cloud strikes create fields which induce voltages in all conductors;

++ ground/earth current flow from nearby cloud-to- ground/earth discharges couples into the supply ground/earthing network via common impedance paths;

++ primary surge arrestor operation or flashover in the internal building wiring causes dV/dt transients.

Fault clearance upstream in the distribution network produces transients with energy proportional to 0.5.L.I 2 trapped in the power system inductance. The energy will depend on the let-through current of the clearance device (fuse or circuit breaker) which can be hundreds of amps in residential or commercial circuits, and higher for some industrial supplies. Power factor correction capacitor switching operations generate damped oscillations at very low frequency (typically kHz) lasting for several hundred microseconds.

The surge test

The test defined in IEC 61000-4-5 is intended to demonstrate the immunity of equipment when subjected to high energy surge voltages from the above effects. The transients are coupled into the power, I/O and telecommunication lines. The surge generator called up in the test has a combination of current and voltage waveforms specified, since protective devices in the EUT (or if they are absent, flashover or component breakdown) will inherently switch from high to low impedance as they operate. The values of the generator's circuit elements are defined so that the generator delivers a 1.2/50~ts voltage surge across a high-resistance load (more than 100 -ohm) and an 8/20~s current surge into a short circuit. These waveforms must be maintained with a coupling/decoupling network in place, but are not specified with the EUT itself connected.

Three different source impedances are also recommended, depending on the application of the test voltage and the expected operating conditions of the EUT. The effective output impedance of the generator itself, defined as the ratio of peak open circuit output voltage to peak short circuit output current, is 2fL Additional resistors of 10 or 40 -ohm are added in series to increase the effective source impedance as specified in the standard.

---The IEC 61000-4-5 surge test

Surge coupling and application

High energy surges are applied to the power port between phases and from phase to ground/earth. For input/output lines, again both line-to-line and line-to- ground/earth surges are applied, but from a higher impedance. 2 -ohm represents the differential source impedance of the power supply network, 12 -ohm represents the line-to- ground/earth power network impedance while 42~2 represents the source impedance both line-to-line and line-to- ground/earth of all other lines.

Power line surges are applied via a coupling/decoupling network incorporating a back filter, which avoids adverse effects on other equipment powered from the same supply, and provides sufficient impedance to allow the surge voltage to be fully developed. For line-to-line coupling the generator output must float, though for line-to- ground/earth coupling it can be ground/earthed. A 10 -ohm resistor is included in series with the output for line-to- ground/earth coupling.

I/O line surges are applied in series with a 40 -ohm resistor either via capacitive coupling with a decoupling filter facing any necessary auxiliary equipment, or by spark-gap coupling where the bandwidth of the I/O line is wide enough that capacitive coupling would affect its operation. Note that testing with differential mode surges is only intended for unbalanced circuits in critical applications, i.e. ports intended for process measurement and control. Long-distance telecommunication lines require special treatment, involving the CCITT 10 x 700~ts waveform applied at up to 4kV. The purpose of the surge immunity test at equipment level is to ensure that the equipment can withstand a specified level of transient interference without failure or upset. In this case it is assumed that the equipment is fitted with surge protection devices (varistors, zeners etc.). Typically such devices have low average power ratings, even though they can dissipate or handle high instantaneous currents or energies. So the maximum repetition rate of applied surges will normally be limited by the capabilities of the devices in use, and a maximum of 10 surges (5 positive and 5 negative) is recommended for any one test procedure. Over-enthusiastic testing may lead to premature and unnecessary damage to the equipment, with possible consequential damage also occurring. Because of this latter risk, some test houses take the precaution of physically isolating the EUT during the test. In any case, the EUT should be disconnected from other equipment where possible and the whole set-up should be well insulated to prevent flashover.

Testing at higher and lower levels

Upsets can occur at some stress levels but not occur at higher stress levels. Testing solely at the maximum stress level specified by the product or generic standard may thus provide a misleading result. A fraction of the maximum test level should be applied at first, increasing until the maximum is achieved. Typically, surge testing starts at 500V and increases in 500V steps. It may also be desired to apply an even higher test level to determine the "margin" that has been achieved, but this should only be undertaken where the possibility of damage to the equipment under test and consequent financial or time losses have been fully evaluated and found acceptable.

Where surge protective devices are fitted in an apparatus, it is advisable that tests are carried out at stress levels just below, and just above, the levels at which the protective devices are intended to operate. This is because actual surges with levels less than the maximum are much more likely, and it is sometimes easier to design a surge protection device that is more effective the higher the surge level. So it has been found in practice that apparatus which meets the maximum surge test levels can suffer field failure with less powerful surges. For example, a gas discharge tube (GDT) device will only trigger when a voltage threshold has been passed. When the maximum surge test level is applied, the GDT only lets-through an overvoltage for a few microseconds before it triggers and protects the circuit. But a lower level of surge voltage may be insufficient to trigger the GDT, allowing the surge to be applied to the circuit for its full duration and possibly causing damage.

Surging in situ

IEC 61000-4-5 does not give any guidance for in-situ tests, in contrast to the previous two standards. This may indicate that such testing is not expected to be commonplace.

The surge waveform has a much lower bandwidth than either the ESD or EFT test pulses, and so radiative coupling is rarely a problem. The major issue is that the surge is a high energy destructive transient, and therefore it is essential to implement decoupling (or back filter) networks to ensure that the energy is only applied to the equipment under test. Trying to apply the surge without these networks onto a power supply that was shared by several systems, would be inviting disaster. (It is also debatable whether a client who has already expended considerable resources to install an expensive system, would want that system deliberately exposed to the potential damage that a surge test can cause.) If the connections to the potential EUT can be broken to insert suitable coupling/decoupling networks, and any ancillary or adjacent apparatus can be properly isolated from the surge, then surge testing in-situ is feasible, with all the above caveats.

Otherwise it is not.

Radiated and conducted RF

The question of radio frequency immunity testing has probably caused more controversy and discussion in the EMC field than any other phenomenon. It is undoubtedly necessary, since there is a proliferation of portable and fixed radio transmitters which can cause field strengths high enough to upset electronic equipment.

It is unfortunately also expensive to do, at least according to the standard methods.

There is thus considerable pressure on test engineers, particularly in manufacturing companies, to find shortcuts in the RF immunity test. Some of these are already half- established and it is worth discussing them in some detail to analyze their advantages and shortcomings.

To do this it is necessary to understand the basic mechanism for RF susceptibility.

Radiated RF fields are developed by all intentional transmitters as a consequence of their function; close in to the transmitters the fields are intense but they decay proportionally to distance. An introduction to the properties of RF fields can be Googled. The fields couple with susceptible equipment by two mechanisms:

++ direct interaction of the field with the circuit wiring and printed circuit traces inside the equipment;

++ induction of common mode currents onto cables, which are then coupled into the equipment by conduction through the cable port interfaces.

Other sections of this guide have described these mechanisms, and particularly how to defeat them, in more detail. They are to a large extent the reciprocal of the coupling mechanisms that are tested in the RF emissions tests. The differences in approach that exist between RF emissions and immunity tests have to do partly with concerns of spectrum management and partly with the historical bias of the relevant standards committees. Reciprocity of coupling can be assumed, because the coupling paths are linear; but the sources and victims of the disturbances are not, and in general they cannot be assumed to be the same, which adds a degree of complication to any effort to harmonize test methods between the two.

Test methods have developed independently for different sectors. The most established standard methods are related to the military/aerospace sector; the automotive sector (ISO/IEC 11452, SAE J 1113); and the commercial sector (IEC 61000-4-3 and -4-6, which developed from the IEC 801 series, originally written for industrial process control and instrumentation). Although there are a few points of contact between these, they are mostly separate, with little commonality in the methods or equipment. They all have in common an approach which applies conducted testing at the lower frequencies and radiated at the higher, although in some cases there is considerable overlap of the ranges. The rationale for this is that cables, being electrically long, are the dominant route for coupling at low frequencies; once the cable becomes much longer than a wavelength (above, say, 200MHz) its efficiency as a coupling path declines, and coupling with the equipment enclosure becomes dominant.

Another factor that these standards have in common is that none of them allow for in-situ testing.

The methods of applying the conducted signal or generating the radiated field are mostly different, but it is generally true to say that developing a uniform, calibrated radiated field is harder at low frequencies than injecting current or voltage onto a cable.

It is the goal of each standard to devise a test method which both simulates the real environment for its particular application sector adequately, and yet allows minimum uncertainties so that the test is repeatable between test houses and at different times, on the same EUT. It is this latter aspect which is hardest to achieve for RF immunity, and which accounts for much of the complexity and expense of the various methods.

This section will look at the standard methods, mostly in the IEC documents, for conducted and radiated tests, and then discuss how these may be applied, where possible, in situ, and what alternatives might exist.

Conducted methods

Disturbances can be induced directly onto cables by two principal methods- voltage coupling, which requires a connection to be made to each conductor, or current coupling, in which a wideband current probe is clamped around the cable, and which requires no direct connection.

Direct voltage injection

This is the preferred method in IEC 61000-4-6, and it employs a specific coupling- decoupling network (CDN). Different CDNs are required for each type of cable, since the coupling is invasive, and this has restricted the use of the technique to commonly- employed cables, particularly power cables. It is particularly unsuitable for unscreened cables carrying wideband data. The CDN is designed to establish a 150 -ohm common mode impedance at the port which is connected to the EUT, and to isolate this port from the end of the cable away from the EUT, i.e. the power supply or ancillary equipment.

---The set-up for a three-core mains cable.

The CDN is similar but not identical to the CDN used for transient coupling, and the AMN/LISN used for conducted emissions tests. The main difference is in the 150~2 impedance, with coupling to the lines in common mode only. The requirement for proper bonding to the ground reference plane is as important as in the other tests.

--- Direct injection set-up according to IEC 61000-4-6 The standard requires a test level to be set up into a 150~2 calibration load, rather than into the EUT itself. The power that is needed to maintain the specified level is then re-played across the frequency range with the EUT in place. The major advantage of the CDN method is the low power required for standard test levels -- a test of 10V emf can be achieved by a power amplifier of 7W rating. CDNs are specified, and can be built, to give a common mode impedance within +20~ of 150 -ohm from 150kHz to 26MHz, and over a wider tolerance up to 80MHz. The test can under certain circumstances be extended up to 230MHz, if properly characterized CDNs are available- most commercial ones do not extend this high.

Bulk current injection

As an alternative to voltage injection, IEC 61000-4-6 allows the use of a current injection probe, over the same frequency range. This method has become known as Bulk Current Injection (BCI). Very similar, but not identical, methods are specified in the military standard DEF STAN 59-41, test DCS02, and the automotive components standard ISO 11452-4. ISO 11451-4, the companion standard for whole vehicle tests, also describes the BCI method. The automotive standards give a frequency range from 1MHz to 400MHz, while the military specification extends from 50kHz to 400MHz.

The BCI probe is simply a current transformer whose secondary is the cable under test, and into whose primary is fed the disturbance signal. It is halved and hinged, so that it can be clamped over the cable harness without having to break into it.

There are two ways of applying the BCI test. Because the EUT's common mode impedance is unknown and frequency-dependent, simply attempting to apply a fixed level of current would result in overtesting at some frequencies and under-testing at others. One method, known as the "substitution method", calibrates the power required to generate the required disturbance current into a fixed impedance calibration jig, and then requires this power profile to be repeated with the EUT in place. This is the only method allowed in IEC 61000-4-6 and DEF STAN 59-41 DCS02. The induced current can be monitored by a separate probe placed between the injection probe and the EUT (), so that it can be limited if it exceeds some higher level than the expected test level, if the EUT impedance drops to a low level at some frequency.

The second method, allowed by ISO 11452-4, is known as the "closed loop" method. In this case the monitor probe records the actual injected current, which is increased at each frequency until either the maximum test current level is reached, or the maximum power to the injection probe is reached.

In all cases the test set-up must be carefully controlled with both the cable under test and the EUT mounted a set height above the ground reference plane. Especially at the higher frequencies, the cable layout and the distance between the probes and the EUT are critical; large variations occur for small changes in probe position. The impedance at the far end of the tested cable has to be stabilized in some way; for the military and automotive tests, it is usual to connect the actual ancillary items mounted on the ground plane as they would be in the real vehicle, on the assumption that this provides a reasonably realistic situation. For tests on commercial products, ancillary equipment is often ill-defined and instead some method of impedance stabilization to 150.0. is required.

Although it requires more power for a defined injection level than direct voltage injection (because of the power lost within the probe), the great merit of the BCI method is that it can be used on any cable that can pass through the aperture of the chosen probe.

This makes it highly suitable for in-situ tests.

--- Bulk current injection set-up

EM-Clamp injection

A similar transducer to the BCI probe is the EM-Clamp, which is the third alternative specified in IEC 61000-4-6. This operates along similar lines but couples both inductively and capacitively at the same time; through the use of ferrite absorber material it allows a certain degree of isolation from the ancillary equipment end of the cable. It is considerably bulkier and much longer than the BCI probe and although it is frequently used in laboratory tests, tends to be inconvenient for in-situ purposes, where the cables to be tested are usually less accessible.

Radiated methods

The basic approach to radiated RF testing is to develop a constant, uniform field over a specific volume in which the EUT is immersed. Provided that the field level is uniform and the EUT's coupling to it is invariant, the assumption is made that the test will repeatably and representatively exercise all the EUT's susceptibilities that need to be tested. Actually achieving an adequately controlled field and coupling is the goal towards which all standard radiated immunity test methods strive.

The most thoroughly specified method of developing the field is from an antenna in a test chamber, and this is described next. Alternatives do exist for equipment tests - the TEM cell, the GTEM cell, or the stripline -- but these have no relevance to large system testing and are not pursued further here. The reverberating chamber method has many fans (sic) and is particularly suited to developing high field levels, but it also has several functional and operational disadvantages and is yet to be specified in any commercial standards, although a draft document is in existence.

Antenna and screened anechoic chamber

IEC 61000-4-3 describes the requirements and methods for a radiated field test from 80MHz to 1GHz using an antenna. After many years in which all standards assumed a closed-loop control method, in which the field at the EUT was monitored and controlled during the test, this approach has finally been scrapped in favor of the substitution method. Exactly as with the conducted immunity substitution method, the field strength is first calibrated over a uniform area without the EUT present; and then the power level which gives this field strength is re-played with the EUT in place and aligned with the uniform area, in the various orientations that have to be investigated ().

Procedure:

1. Calibrate field uniformity without EUT

2. Calibrate required field level, without EUT

3. Re-play power vs. frequency for all orientations and polarizations, with EUT

--- Generating an RF field in an anechoic chamber

The rationale for this method is that the presence of an EUT inevitably distorts the field, especially at frequencies where the EUT is large compared with a wavelength, and so it is impossible to specify a precise applied field strength with the closed loop method. If the monitoring probe is placed in one position next to the EUT, the field at any other position is unknown but is likely to be different. This leads to the possibility of severe over- or under-testing if the probe is positioned at a point which exhibits a field null or peak at any given frequency.

The benefit of the substitution method is that the field strength at any given point around the EUT is irrelevant for the test outcome. What matters is the field strength for which the set-up is calibrated in the EUT's absence; distortions introduced by the EUT are therefore circumvented.

In practice, there are a number of major difficulties with this method, the most important being that it relies on an area of uniform field strength being generated within which the EUT will be placed. The larger the EUT, the larger this area must be. The standard specifies the limits of uniformity to be -0, +6dB over three quarters of the points at which it is measured. It is specified in this asymmetrical way so that under- testing will not occur at any point, but by implication, over-testing of up to 6dB (i.e. 2 times, or 20V/m instead of 10V/m) can still occur. Even with this apparently wide margin, a very expensive test facility is needed. The test must be performed inside a screened chamber to avoid interfering with other radio services. The screened room walls introduce reflections which, if untreated, would create deep nulls in the field pattern over distances of a half wavelength, i.e. 1.875m at 80MHz down to 15cm at 1GHz, which would completely demolish the field uniformity. Anechoic material (carbon foam or ferrite tiles) which is effective over the whole frequency range must line all six surfaces of the chamber, and this greatly increases the facility cost.

But there is more to come. The size of the uniform area also depends on the directivity of the antenna that is used. To maximize the area, the antenna needs to be as far away as possible from the plane in which the field uniformity is measured. This also has the beneficial effect that interaction between the EUT and antenna, which would modify the applied field from that which was achieved in calibration, is minimized. But the further away the antenna is, the more power is needed for a given field strength. The power is actually proportional to the square of the separation distance, so that the power needed for a 3m distance (preferred in the standard) will be nine times that needed at lm. 10V/m at 3m will require typically 100 watts from 80MHz to 1GHz with a conventional BiLog antenna ( ---- note how the power requirement climbs sharply at lower frequencies). A broadband power amplifier covering the same frequency range at this power level is a further major expense.

--- Required power versus frequency for different antennas (source: Schaffner-Chase EMC)

Subsidiary requirements of the test method are:

++ modulation must be applied to the test signal, since real disturbances are likely to be modulated in one way or another, and this affects the EUT's response; the default is 1kHz sinusoidal at 80% modulation depth. This increases the actual peak of the applied level by 1.8 times (18V/m rather than 10V/m) and demands an extra 5.1 dB from the power amplifier over the unmodulated level

++ the test signal must be scanned or stepped over the frequency range of 80- 1000MHz at a rate slow enough for the EUT to respond. For swept measurements the standard requires a sweep rate no faster than 1.5 -3 decades per second. For stepped testing, which is the norm for computer- controlled tests, the step size should be no more than 1% of the previous frequency (i.e. logarithmically increasing with frequency), but there is no explicit requirement for dwell time at each frequency; applying the above sweep rate works out to nearly three seconds per step and an overall minimum time of around 15 minutes per full sweep

++ the EUT must be illuminated with both horizontal and vertical polarization on all four sides, or all six sides if it can be used in any orientation, so that a total of eight or twelve sweeps are required.

The first two points apply also to the conducted test.

In-situ tests

Reading through all the above requirements and methods, you might be forgiven for wondering how they can be applied in situ. The answer, of course, is that they can't, at least for the radiated tests.

The main issue with radiated testing in situ is that it is illegal (in the UK; it may not be in other countries, even in Europe). Since the installation in situ is by definition not inside a screened chamber, any radiated test would be generating unacceptably high levels of disturbance signal which could affect other radio equipment over quite a wide area, certainly exceeding the boundaries of the installation.

It might be possible to apply for a special test license to do this on a site-specific basis, but the grant of such a license would be subject to inevitable bureaucracy (pace the Radio communications Agency) and might contain conditions which would make it unworkable in practice.

High levels of radiated field can also have implications for human health.

Guidelines on acceptable levels of field to which personnel can be exposed are contained in and other documents, which should be consulted if this question arises in an RF immunity testing context.

But some form of RF immunity testing is a necessity if the installation is to be shown to be compliant, either with the immunity requirements of the EMC Directive, or with operational safety and reliability requirements in the case of a contractual EMC obligation. Therefore, some way must be found to get around the legal restriction. There are three main possibilities, which have been explored by competent bodies who have to approve Technical Construction Files, and they tend to be used in combination.

Immunity tests of modules, combined with a coupling analysis

RF immunity testing is a pre-requisite of apparatus that is to be CE marked, and if the results of such tests are available they can be fed into an EMC analysis for the apparatus when it is incorporated into a system or installation. If the apparatus (such as a chart recorder or temperature controller) is to be built into a screened enclosure, the known shielding effectiveness may be used (with caution, i.e. a safety margin of 10-20dB) to justify a reduction in the immunity level for that apparatus if necessary. The usual difficulty with this approach at present is that sufficient information is not available on which to base a sound analysis, or the equipment is supplied without an adequate form of EMC compliance statement-- see section on procurement specifications for a further discussion.

Limited-frequency RF testing using pre-existing licensed transmitters In many cases the main portable transmitter frequencies that will be present in the EM environment on the installation site, and which therefore pose statistically the greatest threat to the system, are known. Examples are:

• ++security and plant engineers' communication transmitters
• ++ cellular mobile phones
• ++ CB and radio amateur transmitters

These transmitters are available to the site operator and are licensed for use on the site, and can therefore be used for testing easily and cheaply without restrictions. Keying such a portable device while holding it near to potentially susceptible parts of the system under test may not be particularly "scientific", but it can be effective at exposing weak areas of the installation, and pointing out the need for remedial measures. The disadvantage is that testing with these devices does not characterize the EUT’s response to other frequencies that may be introduced throughout the lifetime of the installation, for example by the emergency services, and therefore such tests should not be relied upon on their own.

Some effort should be made when applying this form of ad-hoc testing to characterize the field strength generated by the device and applied to the EUT. This can be done quite easily with a broadband field strength meter positioned at various distances and orientations from the transmitter, but beware of errors introduced by near field effects such as interactions with personnel and nearby objects. Also, beware of the intelligence of digital cell-phones, which adapt their transmitted output according to the path loss, which can change very rapidly as the unit is moved around the EUT; checks of the actual emitted field strength, and/or "engineering fixes" to ensure maximum output, are essential with these devices.

Substitution of conducted immunity tests

Because the conducted immunity test methods, particularly BCI, are relatively easy to perform, they can be applied to many cables on the system under test over as wide a frequency range as possible. Typically, the BCI method can be applied down to 50kHz and up to 400MHz as in the military tests if it is felt necessary. Since the cables for the system are already terminated in their actual impedances, the artificial test set-up of the laboratory BCI test need not apply; all that is necessary is for a pre-calibrated probe system to be used on each cable in its actual position that might act as a coupling route for local radiated fields. It is the test engineer's responsibility to select appropriate cables based on knowledge of the signals they carry, the equipment they connect to and their routes between the equipment. The levels of injected current do not correlate directly with radiated field levels; different authorities can quote anywhere between 1 and 10mA as giving about the same stress level as a 1V/m field.

Strictly speaking, though, conducted immunity tests could be as serious a source of radiated disturbances as is a radiated test. The cable on which the currents are induced will act as an antenna, albeit not as efficiently as an intentional one. The legality of performing conducted tests outside a screened room is questionable, and a test engineer who takes the responsibility of doing so should be aware of the potential for affecting legitimate radio users.

One possible way to reduce the magnitude of this problem is to fit a number of clip- on ferrite suppressors to the cable on the "other" side of the BCI probe. The probe should be fitted close to the unit to be tested in any case, and the ferrites should help prevent the rest of the cable from radiating too much. The reduction achieved is unpredictable, due to the influence of other cables and metalwork; the method is also modifying the system configuration, which could be unacceptable for system-wide testing. A broadband omnidirectional field probe could be used to measure the actual fields "leaked" by the test method and the consequent reduction by applying ferrites to the untested cable lengths.

Next:

Prev: Emissions

top of page  Article Index   Home