This page is relevant to testing for both the EMC Directive and the R&TTE Directive. It covers the EMC testing common to both Directives and the Radio testing for the R&TTE Directive. Telecommunications testing is not included as it is relatively straightforward compared to EMC and Radio. Telecommunications testing is mostly concerned with the signal characteristics (e.g. waveform, timing, jitter etc.) and the characteristics of the signal line (e.g. impedance). To a large extent the characteristics are controlled by dedicated chipsets which are designed specifically for the desired type of interface, or can be so configured through firmware. The only problems we have come across were with waveform characteristics, timing jitter and termination impedance degraded by transient protection devices.
In this section we deal with some of the major aspects which can affect the radiated emissions from an EUT. We focus on radiated emissions because it is the area where there is the greatest scope for problems, particularly for plastic cased products, and it is the area which requires the most time consuming and expensive level or re-work to rectify. An equally important area which is just as likely to cause problems is conducted emissions, particularly because of the steady increase in switching frequency of switched mode power supplies. We offer suggested targets for the EMC performance of certain components and sub-assemblies which should help to ensure completed equipment meets its objectives and that the manufacturer is not left with the job of reconciling the poor performance of constituent parts by introducing controls at the design and procurement stage of the project.
The idea is to impose limits on the constituent parts of a piece of equipment whether designed in-house or supplied by an external contractor. This increases the chances of success first time and mitigates the delay and expense of correcting the problem after the fact. Once you have accepted delivery of a product, it is harder to correct and we feel it is better to get suppliers to contribute towards the cost of compliance. This strategy was borne out in practice with one of our customers where no fewer than five systems passed first time with a large margin as a result of the application of margins to sub-assemblies and pre-compliance testing to these margins. There were three subsequent instances where problems arose and all were due to our guidelines being ignored.
|
|
|
|
|
|
| Equipment Enclosure |
|
|
||
|
|
|
|
Only applies if the enclosure contains electronics e.g. conditioning or power supplies. | |
|
|
Margin 10dB(BB), 6dB (NB) |
|
Limits as for final product. BB = broadband NB = narrow band |
|
| Equipment card rack in closed cabinet/enclosure |
|
|
All panels/doors in place and dummy circuit card faceplates. | |
|
|
|
|
||
|
|
|
|
Only applies if there is no power supply or filtering external to the rack between the power input to the rack and the external power input port. | |
| Single Circuit Card in Plastic Enclosure |
|
|
|
Assumes no conductive coating on enclosure but can include local shielding on pcb. |
| Single Circuit Card in Equipment Rack which provides shielding |
|
|
|
Assumes a predominately digital card powered from the rack. |
| RF or Analogue Circuit Card with small digital content |
|
|
|
Assumes a card with a small low power digital content powered from the rack. |
| Any Circuit Card with primary power requirements |
|
|
|
Includes any of the three above where primary power is applied in addition to rack dc power. |
|
Power Converter Unit ac/dc or dc/dc |
|
|
|
Depends on the limit applied to the top level assembly. The preference is for Class B, especially if the PCU is used in quantity. |
|
|
|
|
||
| Enclosed module without primary power |
|
|
|
Assumes metal cased or otherwise shielded powered from system power supplies. |
| Enclosed module with primary power |
|
|
|
|
It should be understood that the performance requirements in the above table are not exhaustive and are just a starting point. It should be possible to tailor the requirements to suit any particular product or system or even generate requirements which are not based on regulatory standards. We have used CISPR 22 instead of EN55022 as it is more likely to be recognised worldwide and we have kept to recognised limits to suit labs which have automated measuring systems. With one of our customers, we generated specific sets of limits and programmed them into the in-house measuring equipment.
As development progresses it may be found necessary to increase limits and this is desirable only if it can be shown that the performance of the final product is not compromised. Equally limits should be restricted if the sub-assembly concerned meets the requirements without cost penalties. Once you have achieved a satisfactory result at the pre-compliance stage for the finished product, it is worthwhile measuring the performance of all the sub-assemblies to establish realistic performance objectives. This is of great advantage if any of the sub-assemblies have multiple vendors as you can freely change suppliers without retesting the product. You can also check the performance after making changes to see if there has been any degradation in (or improvement to) the EMC performance which may require either a design re-think or re-certification of the product. This is particularly true where certification is to the FCC regulatory requirements which require a re-test for any changes or a certification test for all the variants of a product. It is also true in the case of self-certification to both the EMC and R&TTE Directives.
It is however not always appropriate, as in the case of simple products and products with a short market lifetime.
This section applies whether the product comes under the EMC Directive only or the R&TTE Directive. It is important to bear in mind that, when presenting equipment for test, you give yourself the best possible chance of passing without taking special measures which would not otherwise exist and performing modifications which would not be on the production version of the product.
The following precautions should be taken prior to submission for the EMC compliance test:
< Preparation < Design for Compliance < Top
The EMC Test Plan is probably the most important document as it gives the test house all the information it needs to perform the tests on the test sample in a manner which is clearly identifiable and repeatable. In spite of the fact that the EMC standards which define the series of EMC emission and immunity tests are available, there is still a considerable variation in the implementation of the test methods from test house to test house even though they are all probably UKAS accredited. This inevitably leads to variations in the final result which are on top of intrinsic variations due to normal test equipment calibration. Having the test set-ups accurately documented in a test plan should reduce the degree of variation. The minimum content of the test plan should therefore be as follows, not necessarily in the order listed below:
< Test Plan < Preparation < Design for Compliance < Top
Not all EMC tests require comment or precautions as the procedures are well documented in the standards and the tests are relatively straightforward.
When it comes to the test, you should have performed all your pre-certification testing either in your pre-certification facility if you have one, or at an external test facility. You should be about 80-90% confident of passing all the certification testing first time. The confidence level will be higher with the more margin you have from the pre-certification tests. Generally radiated emissions is the problem area, which is why it is normally performed first. It is also the test which is harder to repeat accurately. As the measuring uncertainty is typically ± 4dB, it is a good idea to aim for a margin of at least 6dB with the pre-certification tests.
With the EUT set up for emissions testing and the support equipment set up external to the chamber/OATS, the first step is to perform a 'background' emission test. This is carried out with the support equipment powered up but with the EUT powered down. a test scan should be carried out for 'zero' emissions. The reason for this is that it is the EUT that is on test and not the support equipment. In our experience, test laboratories have tried to claim that if emissions are caused by the interaction between two pieces of equipment then these emissions should be included in the certification tests for the EUT. This is of course not the case. If the support equipment interaction was to be included, it would be inside the chamber as well.
If the result gives significant emissions then the cause must be traced and rectified before proceeding with the certification tests. To be acceptable, background emissions must be less than 6dB below the test limit required, so aim for at least 10dB. It may be necessary to introduce isolation between the EUT and the support equipment by means of filtering. The following is a list of things to try if a problem arises:
It is not normally necessary to populate all the identical ports. The rule of thumb is to add cables until the increase in emissions increases by less than 2dB. This rule should be applied for all the interface types if the EUT has several ports and several different types of interface.
Once you are satisfied that the ambient is zero or sufficiently low and the EUT is operating correctly, the Radiated Emissions test can begin. This is normally a chamber scan with no height scanning primarily to identify possible problem frequencies. The equipment is then transferred to the Open Area Test Site (OATS) for final measurement of any possible problem frequencies so identified. Due to anomalies with all chambers it is not uncommon to see emissions appear from nowhere, particularly at the lower end of the frequency spectrum, below the resonant frequency of the chamber where propagation is less efficient.
A word of warning is necessary where the EUT is essentially dc powered from a plug top mains dc adapter as the impedance characteristics of the adapter can affect the radiated emissions from the EUT as well as the conducted emissions. Refer to the paragraph on plug top power supplies below.
< Radiated Emissions < EMC Testing < Test Plan < Preparation < Design for Compliance < Top
The conducted emissions test usually does not require the same setting up as radiated emissions although it is important to check the ambient with the EUT switched off. If the EUT has more than one power cord, then each must meet the requirements on its own. Each must be connected through an identical LISN. The LISNs not in the measuring path must have 50ohm terminations. If the EUT has an internal mains distribution panel terminated in one power cord, then that power cord is the one that is tested and only one LISN is required. The emissions will however be additive to some extent. There is therefore some advantage in having multiple power cords. Conducted emissions are a bit more predictable than radiated emissions and should give a result fairly close to any pre-certification measurements you may have made. We have achieved ± 1dB in the past between our measurements and a test house. Differences of ± 2dB are typical though, due to basic calibration and differences in the LISNs used. You should have performed pre-certification tests and achieved a margin of at least 6dB.
If the EUT can operate from both single phase and three phase supplies, conducted emissions must be measured under both sets of circumstances.
A word of warning is necessary where the EUT is essentially dc powered from a plug top mains dc adapter. This is especially true if the EUT is double insulated and has no earth connection but can be earthed through the shields of the interconnecting cabling. You should perform the following tests to ensure the compliance is not dependent on configuration:
< Conducted Emissions < Radiated Emissions < EMC Testing < Test Plan < Preparation < Design for Compliance < Top
The immunity tests are grouped together as there is less variation likely due to set up.
The most important factor in Immunity testing is the ability to know when the EUT has malfunctioned. It must be possible to monitor the operation remotely as the tests generally are conducted in a shielded enclosure to avoid interference issues. Usually cameras provide visual information, but if the EUT operation is transparent to the human eye, then other means of monitoring must be used. This usually comes down to either a remote monitor console or a computer or piece of dedicated test equipment. The monitoring equipment must be adequately protected from malfunction, either directly or by introducing a 'buffer' in the interface leads. The buffer can take the form of a filter, the introduction of impedance through ferrite clamps or just by shielding and grounding the monitor signal lead. Where the failure criteria is loss or corruption of data, care should be taken to ensure that the corruption is caused by the EUT and not the test equipment. In the case of data handling, the EUT should merely process the data from external test equipment which also checks its integrity (data loop back).
It is not normal to permit power cycling as a means of recovering from malfunction after a disturbance signal or electromagnetic field. Product standards often state the requirements, but in some circumstances it is left to the manufacturer. In these cases it is what can reasonably be expected from that type of equipment. The test lab. may query the criteria if they feel it is not reasonable to expect them from that type of equipment.
The tests and limits will be defined in the standard or be able to be determined by the manufacturer from a range of options.
The stress level can be either 3V/m or 10V/m modulated 80% at 1kHz although some standards require 20V/m. In addition there are pulsed tests to simulate interfering fields from mobile and cordless phones. Usually a slight over test to compensate for measuring inaccuracy is used. The test is carried out in an anechoic or a semi-anechoic chamber to ensure a uniform field and absorbing material is placed between the EUT and the antenna to cut down reflections from the floor of a semi-anechoic chamber. The distance from EUT to antenna is typically 3m. The EUT is illuminated on all four faces and in both polarisations. At the discretion of the test lab., some faces may be omitted on the grounds of symmetry, because the sample is small or very well shielded on certain faces, or because there is no perceived threat for whatever reason. This results in cost savings through reduced test time.
The stress level can be either 3V rms or 10V rms modulated 80% at 1kHz. Usually a slight over test to compensate for measuring inaccuracy is used. The test is carried out in an anechoic or a semi-anechoic chamber because of potential interference and the need for a uniform ground plane. The tests are based on the assumption that a cable set to the specified height above a ground plane has a characteristic impedance of 150ohms. The injection networks are designed to terminate the cable at the source and load ends with this impedance and to insert a high impedance buffer to isolate support equipment. The 50ohm source impedance of the generator is included in the 150ohm termination. The tests and injection methods depend on the cable type used for the signal interface and the following options are typical:
In the event of a failure, it is recommended that investigations are carried out to ascertain the frequencies and levels at which the failures occur and, if time permits, what the thresholds are before a failure occurs even if the EUT passes the stress levels imposed by the standard.
As for the other immunity tests, it is important to lay down the criteria for failure of the EUT and to document it in the Test Plan. Points for the application of the ESD disturbances should be identified in the Test Plan unless you are happy for the test lab. to select them using their own experience and discretion. Ensure, however, that any connectors used only for test purposes are covered to protect the exposed pins. Although not essential, it is a good idea to select ESD points which are behind access covers or doors. This covers the situation where maintenance or service personnel come into contact with the equipment without first employing anti-static measures. The US Telecomms standard for example, Telcordia GR-1089-CORE, requires full functionality of telecommunications equipment during an ESD event of 15kV (air discharge) and 8kV (contact discharge) with all doors and access panels open.
This test is applied common mode to ac and dc power by means of a coupling network, and to signal and control cables by means of a capacitive clamp. It is generally only necessary to test one of each of the types of signal interfaces unless the EUT has a rack system where there can be differences in grounding (e.g. front panels of plug-in circuit cards). On the other hand, all the cables of the same type or group can be tested at the same time, within limits of the capacity of the clamp. It is a good idea to 'buffer' the remote ends of the signal and control cables with ferrite clamps to reduce interference to support equipment. There are different levels of stress for this test which should be stated in the product standard. Typically the stress levels are 1 or 2kV peak for ac power and either 500V or 1kV peak for signal and control cables depending on the application and the environment.
< Immunity Tests < Conducted Emissions < Radiated Emissions < EMC Testing < Test Plan < Preparation < Design for Compliance < Top
Radio testing is divided into two areas, the testing of the radio itself and the testing of ancillary equipment associated with the radio. Ancillary equipment is any equipment which the radio requires to operate as intended. The radio and ancillary equipment may be in separate enclosures or may be contained in the same enclosure. Equally the radio may be capable of installation in more than one enclosure containing ancillary equipment. A Test Plan for both the radio and ancillary equipment should be prepared. For both radio testing and the EMC testing of the ancillary equipment, the general preparation should follow the same guidelines as for EMC testing above. This is particularly true if the radio is housed in the same enclosure as as the ancillary equipment as emission and immunity tests will be performed on the enclosure in addition to the radio performance tests.
The EMC Test Plan should follow the same guidelines for the radio+ancillary/ancillary equipment as for the general EMC case above. Where the radio is contained within the same enclosure as the ancillary equipment, both the radio tests and the EMC tests should be carried out with the radio installed in the enclosure.
Some of these recommendations may seem obvious but we have seen instances where EMC tests were carried out on a radio/ancillary equipment system by one person while radio tests were being carried out at a different test facility using a second piece of equipment by a different person. The EMC sample had been properly prepared for EMC whereas the radio sample had not. The EMC sample passed with virtually no emissions whereas the radio sample almost failed its spurious emissions tests (EMC Radiated Emissions) through poor preparation.
< Immunity Tests < Conducted Emissions < Radiated Emissions < EMC Testing < Test Plan < Preparation < Design for Compliance < Top
Because of the vast range of different types of radio equipment it is impossible to include everything. Most types will be covered by ETSI product standards and will be either of the integral antenna or removable antenna types. Integral antenna radios are tested in an anechoic chamber using a second calibrated receiving antenna to measure the field strength, and hence the transmitted power (ERP or EIRP) and characteristics of the spectral mask. Measurement is also made of the spurious radiation outside the operating band of the radio.
Testing of non-integral antenna radios can be carried out in a standard lab environment as the power is absorbed in a matched load. Testing of spurious harmonics, spectral mask and power are made using this conducted method. Measurement of spurious radiated power in the operational band is not normally required but is useful for exposure reasons and also to identify sources of leakage which represents a loss of expensively generated rf energy. The power again is measured in terms of ERP or EIRP.
The ERP or EIRP is measured by the substitution method where the field strength is measured at each of the relevant frequencies. The EUT is then replaced by an antenna of known antenna factor and power applied at the same frequency until the previously measured field strength is reached. The desired result is the power supplied by the rf generator. This process is very time consuming and not suitable for the spurious emissions outside the band required by the radio specification - i.e. EMC spurious radiated emissions.
For out of band spurious emissions, a quicker method is to measure the field strength as normal but to replace the normal emission limits from EN55022 with a limit in dBµV/m calculated from the power limit defined in the specification (ERP or EIRP(dBm)). Due regard must be made of the measuring bandwidth required which will be typically 120kHz up to 1GHz and 1MHz from 1GHz to the highest frequency required.
There is also a requirement to measure the radiated spurious emissions outside the band defined by the spectral mask, i.e. from 30MHz to the highest frequency but respecting an exclusion band around the operating frequencies defined by the mask.
Additional testing for radiation hazard and SAR may also be required depending on the product and the specification.
< Radio Preparation < Immunity < Conducted Emissions < Radiated Emissions < EMC Testing < Test Plan < Preparation < Design for Compliance < Top
The testing of ancillary equipment should be included as part of the product standard for the radio. Essentially normal EMC provisions apply with the level set at typically EN55022 Class B for emissions. The method of measurement is the same as EN55022 for both conducted and radiated emissions. The ancillary equipment is therefore treated as though it was a normal piece of equipment without a radio. The treatment is the same for equipment where the radio and ancillary equipment are in the same enclosure. The immunity requirements would normally only apply to the ancillary equipment unless the radio has a removable antenna and is not powered from the ancillary equipment. In such cases immunity would also apply to the radio with perhaps a surge requirement for the antenna port.
Page last updated: 28 October 2007 17:28:03