Standards and Certification
The Development of CISPR 22 as the Basic International EMI Standard for Information Technology Equipment
by Werner Schaefer
Aug 1, 2007

See this article in our Digital Edition
Download a pdf of this article (right click to save the pdf)

CISPR 22 has been used globally for many years to determine compliance of information technology (IT) equipment with applicable limits. Many economies like the European Union, Japan, Australia and New Zealand have adopted CISPR 22, sometimes with modifications, into locally applicable standards. This article will highlight the development of this important standard, and provide a brief preview of future changes.

Background
The first edition of CISPR 22 was published in 1985. This standard was prepared by Subcommittee G of CISPR, an International Special Committee on Radio Interference within the Electrotechnical Commission (IEC). This subcommittee was responsible for the definition of limits and specifics of the test procedures applicable to information technology equipment (ITE). This document has been the dominant international standard for radiated- and conducted-emissions testing for ITE. It has moved through five editions since it was first written in 1985, in order to address new technologies with new setups for equipment under test (EUT) and new measurement methods. Several amendments to these individual editions have been published that contain specific modifications, which have been used in conjunction with the various editions and have typically been incorporated into the next formal edition.

In 2001, a new CISPR subcommittee, CISPR I, was formed out of the former subcommittees CISPR E (responsible for sound and broadcast receivers and associated equipment) and CISPR G. CISPR I was formed in order to better address new technologies, especially in the multi-media area. CISPR I prepared all documents subsequent to CISPR 22, Edition 3 Amendment 1. The current version of the standard, CISPR 22 Ed. 5.2, was published in March 2006 and further changes are already under consideration.

Differences Between the First (1985) and Second (1993) Editions of CISPR 22 [1]
The major changes between the first and second edition of CISPR 22 are summarized in this section. Editorial changes are not included in this comparison.

The structure of the two documents differs significantly. For example, the second edition adds separate sections for scope, normative references and definitions, incorporating some of the material contained in the preface of the first edition. Along with the changes to the structure of the document, the clause numbering was changed to include the new amendments. In general, material after clause 3 in the first edition appears in a section of the second edition numerically increased by one.

The frequency range of interest for the standard was increased in the scope to cover the range from 9 KHz–400 GHz. It was also specified that no limits apply outside this range. It should be noted that no language was included in subclause 4.2 of the second edition explicitly stating that no conducted emissions limits apply in the frequency range from 9 to 150 kHz. Furthermore, no specific mention of the frequency range from 0.15 to 30 MHz was made in clause 5 to explicitly state that no radiated emissions limits apply in this frequency range. However, the exclusions of the frequency ranges mentioned are implicitly documented, since the applicable frequency ranges for the conducted and radiated emission limits were specifically stated.

The clause for normative references was amended, with additional references and updates of existing references. For example, a reference to CISPR 16 (1975) in the first edition, was updated to the 1987 version in the second edition, and a reference to CCITT V.24 (1993) was added.

In addition, several definitions were clarified or updated. For example, the definition of “ITE” was expanded, and “EUT” (equipment under test) was used in the second edition instead of the term “test unit.” “Terminal interference” was changed to “conducted disturbance” and “radiated interference” became “radiated disturbance.”

Wording in clause 7, interpreting RFI limits relating to the significance of limits, tests required, and statistical compliance (the 80/80 rule), was put into clause 8 and became a subclause on “Interpretation of CISPR radio disturbance limit.” Specifically, reference was made to the term “equipment” instead of the previously used term “appliance”.

The quasi-peak radiated emissions limits for 30 m measurements of Class A equipment (Table III of clause 5 in Ed. 1) was changed and included in Table 3 of clause 6.

The measuring distance for Class A radiated emissions measurements was changed from 30 m to 10 m, and the limits were proportionally changed by approximately the amount (10 dB) calculated by the difference in the distances and assuming a 1/r E-field relationship. This change also affected subclause 11.3.6 of the second edition, changing the distance for user installation testing to 10 m. For small products, where the 1/r E-field relationship is valid over the 10–30-m range, this change makes the measurement more convenient, but does not significantly change the requirement burden.

In subclause 10.2, “Artificial Mains Network” (AMN), of the second edition, the following new paragraph appears: “Conducted disturbance is measured between the phase lead and the reference ground and between the neutral lead and the reference ground. Both measured values shall be within the appropriate limits.” Although not included in the first edition, this description was probably intended but not explicitly stated. Therefore, this material was included in the second edition to provide clarity rather than to describe a significant change.

In subclause 11.4, “Measurement in the presence of high ambient signals,” of the second edition, a new alternative test method was introduced that was not in the first edition. The new approach stated: “In the frequency bands where the ambient noise values of clause 9 are exceeded (measured values higher than 6 dB below the limit), the disturbance values of the EUT may be interpolated from the adjacent disturbance values. The interpolated value shall lie on the curve describing a continuous function of the disturbance values adjacent to the ambient noise.” This is not a mandatory requirement but is stated as an advisory comment or helpful note. It merely expands upon situations described in the first edition.

Comparing the Second (1993) and Third (1997) Editions of CISPR 22 [1]
The third edition of CISPR 22 contained much more detailed descriptions of requirements than the second edition. New annexes were added, and coverage of telecommunications equipment testing was expanded. In the third edition, telecom equipment was regarded as a type of information technology equipment, subject to certain conducted emissions tests. Many procedures in the second edition were amended or clarified, providing more details in regard to the test methodology and reducing interpretive variability. Most editorial, non-technical, and other revisions that did not significantly affect substance, are not included in the following comparison. References to the affected clauses, figures, or tables are made to assist readers who need to determine the exact wording differences.

Normative annexes B and C that are part of the requirements of the standard were new in the third edition. Annex B covered the decision tree for peak detector measurements, and Annex C addressed possible test setups for common-mode measurements. Informative annexes D and E were also added for guidance and further explanation. Annex D described the schematic diagrams of impedance stabilization networks, (ISNs) and Annex E covered parameters of signals at telecommunications ports.

The limit for disturbance power measurements that were included in the second edition, clause 7, were deleted (this limit was applicable to absorbing clamp measurements that were used before in certain countries).

EUT Configuration
The EUT configuration during testing was significantly revised (subclauses 8.1 and 10.4 in the third edition). The requirement to maximize the emissions by manipulating cables or equipment configuration was eliminated in favor of configurations that were “consistent with typical applications.” The need to document the configuration such that the test results can be reproduced was retained. Procedures for determining typical configurations for modular equipment or systems consisting of several pieces of equipment were clarified. Detailed examples of configurations for PCs, their peripherals and point-of-sale terminals, were added. Specific requirements for testing separately marketed PC boards (e.g., PC adapter cards or the equivalent) in representative hosts, were also added.

A new subclause 8.1.1 discussed the determination of maximum emission configurations for both conducted and radiated measurements. This subclause referenced 11 drawings that were introduced in the third edition
(Figures 4–14), describing typical initial equipment and cable placement. These drawings, one each for conducted disturbance (subclause 9.3) and radiated disturbance (subclause 10.4), were also referenced in a new equipment setup subclause. This subclause now specified separation distances between equipment and both vertical and horizontal ground reference planes.

EUT Operation
EUT operation, during both conducted and radiated measurements, was detailed in a new subclause 8.2 in the third edition that included guidance on computer test programs, rules for operating visual display devices, and parameters for the operation of facsimile devices and telephone sets. Again, the fundamental requirement was intended for typical operation, but EUT exercising had to “permit detection of all system disturbances” and “generate the greatest level of emission” or “full disturbance potential.”

Conducted-Disturbance Measurement
In clause 9 of the third edition that covered the measurement of conducted disturbances, a recommendation was added that measurements are to be performed in a screened enclosure. Also, the use of the peak detector as an alternative to the quasi-peak (QP) or average detectors was permitted to reduce test time. However, in cases of dispute, the QP or average detector methods and limits took precedence. The decision tree was detailed in the normative Annex B.

Telecom Port Disturbance
New conducted emissions measurements on telecommunications ports were added to the third edition (Tables 3 and 4 in clause 5, subclause 9.5, and Annexes C, D, and E). Telecom ports were defined as ports intended for connection to telecommunications networks, which included the public telephone network and local area networks. Conducted measurements were limited to common mode emissions. Since this requirement was new at the time of the introduction, these measurements required a careful study of the applicable limits, ISN specifications and related annexes.

Each telecom port had to be connected to an ISN, typical auxiliary equipment, or an appropriate simulator. In case a simulator was used, the actual common mode impedance represented had to be measured and stated in the test report. Annex D included general performance parameters for suitable ISNs (subclause 9.5.2 of the third edition). Schematic diagrams were also provided in the annex which represented suggested (but not mandatory) designs. Two alternative methods for compliance testing using ISNs were described (subclauses 9.5.1.1 and 9.5.1.2), with the second alternative designated for Category 3 and Category 5 data cables. The principal difference between these alternatives appeared to be the longitudinal (i.e., common mode) conversion loss specified for the ISNs. When ISNs were not suitable to perform the test, other alternative common-mode measurement methods were given in normative Annex C.

New telecom port disturbance limits were provided in Tables 3 and 4 and had to be met in addition to previously existing limits for mains (AC) ports (see Tables 1 and 2). To comply, equipment with telecom ports could meet either of the specified voltage and current limits for asymmetric-mode disturbance, except that both limits had to be met when using a newly introduced current and voltage probe method described in Annex C (subclause C.1.3). Guidelines for test traffic levels were provided in subclause 9.5.3.

Test Reports
Test reports only had to include the six highest conducted and radiated emission levels of all EUT emissions within 20 dB of the limits (subclauses 9.6 and 10.5 of the third edition).

Multiple Power Cords
For EUTs with multiple power cords, wording was added in subclause 9.2e in the third edition, specifying that line cords not under test may be connected to a multiple outlet, which must then be connected to another artificial mains network (AMN) and had to be separate from the AMN used for the line cord under test.

Open-Area Test Sites and Screened Enclosures
To clarify conducted emissions measurements at open-area test sites or screened enclosures, a sentence was added to subclause 9.3, “Ground Planes,” of the third edition, specifying the distance to the horizontal ground plane as
0.4 m. Otherwise, the wording in this section was identical in both editions. (The 0.4 m distance in CISPR 22 varied from the distance required by the test methods in ANSI C63.4:1992 at the time. The 0.4 m distance harmonized CISPR 22 with other CISPR standards that also specify this distance.)

Equipment Requirements
Equipment requirements in the third edition referred to later versions of CISPR 16. For example, rather than referencing CISPR 16, the third edition required measuring receivers to operate in accordance with CISPR 16-1 and 16-2. Similarly, the third edition required the use of AMNs defined in CISPR 16-1, subclause 11.3. The descriptions of the AMN and EUT configuration were identical in both editions, with clarity provided by the new EUT configuration drawings (see “EUT configuration,” above).

For radiated-disturbance measurements, EMI receivers with peak detectors and IF filter passbands in accordance with Clauses 2 and 3 of CISPR 16-1 were permitted instead of QP detectors (see introduction to clause 10 of the third edition). However, in cases of dispute, QP measurements took precedence.

Radiated Emissions Test Setup
Amendment 1 to the third edition, which was published in August 2000, introduced the mandatory use of ferrite clamps on all cables leaving the table-top EUT for a connection outside the test site (e.g., mains cables, telephone lines, connections to auxiliary equipment located outside the test area). These cables had to be fitted with ferrite clamps placed on the floor at the point where the cable reaches the floor. Only one cable per ferrite clamp was allowed. Figure 10 was updated in accordance with this new requirement.

Multifunction Equipment Operation
Amendment 2 to the third edition, which was published in October 2002, provided an improved definition of the terms “telecommunications/network port” and “Multifunction equipment.” In addition, further details on the operation of multifunction equipment during the test process were provided.

These additions were intended to clarify the previously provided requirements that are applicable particularly to multi-media equipment.

Comparing the Third (1997) and Fourth (2003) Editions of CISPR 22
The fourth edition of CISPR 22, published in April 2003, focused on the conducted emissions measurements on telecommunications ports and provided much more detailed specifications of the longitudinal insertion loss (LCL) of impedance stabilization networks (ISNs) that were to be used for these types of measurements. A new informative annex D was added to provide schematic diagram examples of ISNs as background information. The major changes in the fourth edition of CISPR 22 were as follows:

Relaxation in Table 4
Table 4 documented the limits of conducted common mode (asymmetric mode) disturbance at telecommunication ports in the frequency range 0,15 MHz to 30 MHz for class B equipment. In the third edition, a note 3 to this table was added which provided a provisional relaxation of 10 dB over the frequency range of 6 MHz to 30 MHz for high-speed services that had a significant spectral density in this band. This relaxation was restricted to the common mode disturbance converted by the cable from the wanted signal. This provisional relaxation was removed in the fourth edition of CISPR 22.

Methods of Conformance Testing of Disturbances at Telecommunication Ports
The previous two measurement method alternatives that were specified in subclause 9.5.1.1 and 9.5.1.2 in the third edition were combined in the fourth edition. The references to the different cable categories were eliminated, and a much more detailed description was included in the subsequent subclauses that specify the LCL for the various cable types.

ISN Specifications
The definition of the LCL of various ISNs was redefined using different equations for the various cable types. The specifications shown in Table 1 of this article were adopted in subclause 9.5.2.

Cable Type	Frequency Range	LCL Range
Category 6 (or better) unscreened balanced pair cables	< 2 MHz	±3 dB
	2 MHz to 30 MHz	-3 dB to +6 dB
Category 5 (or better) unscreened balanced pair cables	< 2 MHz	±3 dB
	2 MHz to 30 MHz	-3 dB to +4.5 dB
Category 3 (or better) unscreened balanced cables	150 kHz to 30 MHz	±3 dB
Poorly balanced cables	150 kHz to 30 MHz	±3 dB

Table 1

Furthermore, a detailed description of the voltage division factor was included in this subclause.

Informative Annex D
This new annex provided schematic diagrams showing examples of impedance stabilization networks. ISN block diagrams for the following applications were included:

• ISN for use with unscreened single balanced pairs;
• ISN with high longitudinal conversion loss (LCL) for use with either one or two unscreened balanced pairs;
• ISN with high longitudinal conversion loss (LCL) for use with up to four unscreened balanced pairs;
• ISN, including a 50 Ω source matching network at the voltage measuring port, for use with two unscreened balanced pairs;
• ISN for use with two unscreened balanced pairs;
• ISN, including a 50 Ω source matching network at the voltage measuring port, for use with four unscreened balanced pairs;
• ISN for use with four unscreened balanced pairs;
• ISN for use with coaxial cables, employing an internal common mode choke created by miniature coaxial cable (miniature semi-rigid solid copper screen or miniature double-braided screen coaxial cable) wound on ferrite toroids;
• ISN for use with multi-conductor screened cables, employing an internal common mode choke created by bifilar winding multiple insulated signal wires and an insulated screen-conductor wire on a common magnetic core;
• ISN for use with multi-conductor screened cables, employing an internal common mode choke created by winding a multi-conductor screened cable on ferrite toroid.

1. Scott Roleson, “A Comparison of the First Three Editions of CISPR 22”, Compliance Engineering Magazine, November/December 2000.