This third and final part looks at Section 6 (Interference Potential of EDP/OE), Section 7 (The Commercial EDP/OE Interference Models), Section 8 (Emanation Limits for EDP/OE Products), Section 9 (Comparisons of Recommended Limits with Others), Section 10 (Emanation Measurement), and Section 11 (Conclusions).
In the middle of the 1970s, the United States Federal Communications Commission (FCC) began to look seriously at electromagnetic emissions from electronic data processing (EDP) equipment and office equipment (OE). This growing awareness on the part of the United States telecommunications regulatory body was a result of the increasing number of computers being used by society and the heightened potential to licensed broadcast services due to the proliferation of small electronic-computer sources. The Computer and Business Equipment Manufacturers Association (CBEMA) formed a technical subcommittee to assist in preparing an industry response to the concerns of the FCC.
The first part of this series of articles reviewed the first one-third of the report, including the Title of the Paper, Background, Members of the CBEMA Subcommittee, Table of Contents, Scope, Definitions, Introduction and Section 4 (EDP and OE as a Source of Electromagnetic Emanations). The second part of this review looked at Section 5 (Susceptibility of Communications Receivers to Commercial EDP/OE Emanations) of the report.
Title of the Paper
The title of the published paper was “Limits and Methods of Measurement of Electromagnetic Emanations from Electronic Data Processing and Office Equipment.” The report was prepared by CBEMA Subcommittee 5 on Electromagnetic Interference.
Table of Contents
The report was 183 pages in length. It included a Title page, Foreword, Table of Contents, Scope, Definitions, Introduction, seven major sections, Conclusion, and one Appendix.
Section 6 – Interference Potential of EDP/OE
This section of the report completes the investigation of the paper’s interference model by examining the propagation of electromagnetic emanations from EDP/OE to communication receivers. The examination is primarily empirical and considers the proximity of the EDP/OE to the population of receivers. The report distinguished between commercial products and domestic (used in the home) products.
An analysis of the number of receivers quickly led to the conclusion that AM radio receivers and VHF/UHF Television were the most likely to have interference issues with EDP/OE. One of the report’s footnotes says: ‘There were 178 million AM receivers and 102 Million TV sets in the USA in 1971.”
Because there were very few receivers in the frequency range below 450 kHz, the interference potential of the sub450 kHz receivers was considered to be minimal and the low end of the conducted emission limits was set at 450 kHz. Again, because there were very few receivers above 1000MHz and because, in 1977, the harmonics from clocks in the computer equipment were not prevalent above 300 MHz; it was decided to select 1000 MHz as the high-end limit of the radiated emission limits.
Thirty MHz (30 MHz) was selected as the “break point” between conducted and radiated emissions based on (1) propagation factors, (2) measurement practices, and (3) traditional limit setting. A study of the propagation factors of radiated electromagnetic fields below 30 MHz led the study to say: “Therefore, at typical receiver/product separation distances, direct radiation is expected to be much lower than radiation from the receiver’s power cord caused by conductor emanations.”
The paper did a study of receiving antennas within 100 meters of computer system installations in the USA and Canada. The study covered 243 commercial EDP/OE installations and it observed 826 functional antennas. Furthermore, it was concluded that “89 percent of receiving antennas found within 100 meters of commercial EDP/OE installations can be expected to be 30 meters or more from the installation.” The study concluded a 30-meter horizontal distance was a reasonable control distance. A 10-meter height for the antenna was picked as an “average” antenna height.
Five coupling modes were identified for the model.
Mode 1 Power Cord Radiation of Conducted Emanations
Mode 2 Power Line Radiation of Remotely Generated Conducted Emanations
Mode 3 Noise Source/Cord Radiations < 30 MHz
Mode 4 Internal Radiation or Conducted Emanations
Mode 5 Noise Source/Cord Radiations > 30 MHz
These modes are shown in Figure 1 (Figure 6-1 in Report).
Figure 1: Five EMI propogation/coupling modes
The principal coupling mode in the 30 to 1000 MHz range for radiated emissions is Mode 5, where the emanations are transmitted to the receiving antenna through the air. Attenuation of the electromagnetic fields in the 30-300 MHz range was studied and found to be 23 dB per decade, which agreed closely with the theoretical value of 20 dB. Attenuation of the EM energy through building walls was selected to be 8 dB based on theory and experimental measurements.
Conducted propagation losses depend heavily on the source impedance, power line impedance, and receiver impedance. The model being developed by the study assumed the computer system and the “receiver” both were operating from different metered utility services. Theoretical investigations and field investigations led to a 25-55 dB loss for a common service entrance and 42-89 dB loss for separate service entrances. The minimum values of 25 dB (business/apartment) and 42 dB (business/house) were selected for the EMI model.
Section 7 – The Commercial EDP/OE Interference Models
The model for commercial EDP/OE interference is shown in Figure 2 (Figure 7-1 in Report).
Figure 2: Commercial EDP/OE interference model
The model shows (1) the separation distance for radiated emissions between the commercial source of EDP/OE and the receiving antennas is greater than 30 meters and (2) the conducted emanations have at least two sets of power panels/watt-hour meters between the source and the receiver. Eleven assumptions were outlined in the paper and, in general, the assumptions were all worst case.
The maximum permissible field strength of radiated interference at a receiver located 30 meters from an EDP/OE is given by:
E30 = ES – S/N + AB
E30 = radiated emanation limit for commercial EDP/OE separated by 30 meters or more from a receiver in the selected communication service band (dBuV/m or dBuV/MHz/m)
S/N = receiver signal-to-noise ratio for the selected communications service (dB)
AB = building attenuation factor for the EDP/OE environment (dB)
ES = expected receiver signal strength (dBuV/m).
The equation for computing the limit for conducted emissions from commercial computers is given as:
V50 = V”50 + AL + ES – ET
V50 = conducted emanation limit at desired field strength (dBuV) or (dBuV/MHz)
V”50 = equivalent audio threshold of detectable interference at test condition field strength (dBuV or dbuV/MHz)
AL = propagation loss factor (dB)
ES = desired field strength (dBuV/m)
ET = test condition field strength (dBuV/m).
The conducted equation gives the maximum permissible power line terminal voltage across a fifty-ohm grounded impedance that would result from conducted emissions from commercial EDP/OE separated by two sets of power panels and associated utility hardware.
Section 8 – Emanation Limits for EDP/OE Products
This section concentrates on the calculation of maximum permissible amplitudes for both radiated and conducted emissions from computer and business equipment in the frequency range 450 kHz to 1000 MHz.
Table 1 (Table 8-1 in Report) shows the calculated narrowband and broadband conducted limits in the frequency range 450 kHz to 1.6 MHz (the AM radio band) based on the earlier models and the received field strength protection levels presented earlier in the paper.
Table 1: Calculation matrix for EDP/OE emanation limits (power line conducted) for the 450 kHz to 1.6 MHz range
Because of the lack of receivers in the 1.6 MHz to 30 MHz range and the fact that the limit developed for AM radios was conservative, the paper proposed a more relaxed limit of 10 dB in the higher conducted emission frequency range.
Figure 3 (Figure 8-1 in Report) shows the calculated conducted limits for a 50-ohm LISN.
Figure 3: Calculated conducted limits for 50 ohm LISN
For the calculation of radiated emission limits in the frequency range from 30MHz to 1000 MHz, a typical received field strength for a primary service was selected. Then this field strength was decreased by the mean S/N ratio and increased by the worst-case building attenuation factor. A basic limit was calculated for each major communication service for a 30-meter control distance using this technique. The results are summarized in Table 2 (Table 8-2 in Report).
Table 2: Calculation matrix for EDP/OE emanation limits (radiated 30 meters) for the 30–1000 MHz range
The above values were smoothed to (1) produce a regular-unified limit and (2) to allow a single quasi-peak limit for emissions expected to produce narrowband or broadband receiver response.
The “smoothed” limits are shown in Figure 4 (Figure 8-3 in Report).
Figure 4: Recommended emanation limits (radiated) for EDP/OE in the 30-1000 MHz frequency range when measured at a 30-meter antenna distance
The report went into additional details and discussion on the amount of “underprotection” and “overprotection” provided by the “smoothed” limits. Potential relaxation of limits relative to pulse repetition rates for broadband peak limits was also discussed in detail, and then the report concluded that “the most efficient method for obtaining relaxation with calculating PRR is to use a CISPR Quasi-Peak receiver.”
The probability for conducted EMI disturbance was found to be 0.004 for the business/apartment model and 0.0002 for the house/house model.
Section 9.0 – Comparison of Recommended Limits with Others
A study of existing emission limits was performed. With the exception of a European Computer Manufacturer’s Association (ECMA) study published in March of 1976, no requirements and/or limits were found which were specifically developed for computers and office equipment. The technical derivations of national and foreign limits for equipment were not available to the developers of the CBEMA study. Also, the background on the degree of protection provided to communication services by meeting the existing foreign and national limits was not known.
Despite these weaknesses, the CBEMA limits were compared to the existing radiated emission limits and also to the typical NB emanations from existing EDP/OE products. In general, the proposed limits were at the same level as the German High Frequency Law in effect in 1977 and the proposed ECMA limits in the TV and FM tadio bands, and more conservative by approximately 20 dB in other ranges. However, the proposed FCC limits for restricted radiation devices, as per FCC Docket 20780, were more stringent by 10 to 20 dB than the proposed CBEMA limits.
This is illustrated in Figure 5 (Figure 9-1 in Report).
Figure 5: CBEMA recommended radiation emission limits compared to those existing in 1977
A similar analysis was done for broadband limits. The proposed CBEMA limits are shown in the report to be (1) within 2 to 5 dB of the normalized German VDE limit, (2) within 3 dB of the ECMA proposed limits for the European TV and FM radio services, and (3) 3-27 dB more conservative in the other frequency bands.
The conducted emanation limits proposed by CBEMA were then compared to the existing conducted emission limits from ECMA and the German VDE organization. The VDE limit could be more than 12 dB more liberal to more than 25 dB more stringent. The proposed FCC limit on commercial EDP/OE was 14 to 30 dB more stringent than the proposed CBEMA limits. These comparisons
are shown in Figure 6 (Figure 9-4 in Report).
Figure 6: Comparison of conducted emanation narrowband limits
Again, a similar analysis was done for broadband conducted emanations. Some difficulties were experienced in the BB comparison due to the use of a 150-ohm network by the VDE in Germany and because the BB Threshold of Detectable Interference (TDI) was 21 dB greater than the NB TDI. In general, the German VDE limits were more stringent than the proposed CBEMA BB limits.
A comparison of the proposed CBEMA limits and the existing emanation spectra from about 135 products representing the current line of EDP/OE products was performed. A comparison of the data showed that a significant number of installed products had exceeded the proposed limits, but there had been a negligible interference rate with the EDP/OE products. Also, it showed that many of the products far exceed the proposed FCC limits.
This is illustrated in Figure 7 (Figure 9-7 in Report).
Figure 7: Narrowband radiated emanations at 30 meters
A similar graph was generated with similar results for narrowband conducted emissions with a 50-ohm LISN.
Section 10 – Emanation Measurement
This section of the report recommends a standard way of measuring emissions with standardized test equipment. The recommendations included the following:
- tuned RF voltmeters with a CISPR quasipeak detector and/or a peak and average detector
- 50-ohm Line Impedance Stabilization Network (LISN)
- free-field radiated emission measurement or equivalent with allowance for practical site characteristics
- permissible measurement at antenna distances other than the control distance of 30 meters and down to 3 meters
- measurement of products with only normal operational ground configuration; i.e., grounded or ungrounded
- measurement of either units or systems at the manufacturer’s option
The study endorsed the CISPR bandwidth specifications in the 30-1000 MHz range (120 kHz bandwidth) but thought that the 9 kHz bandwidth of the CISPR receiver for the 150 kHz to 30 MHz range was being too stringent by about 20 dB. Fifty-ohm measuring receivers are preferred by the study, but other input impedances are acceptable provided the impedance of the 50-ohm LISN was maintained.
For radiated EDP/OE emanations, the NB limit is equal to the QP broadband limit, so a single QP measurement can be made to evaluate a source of emission that will produce both NB and BB responses. For conducted EDP/OE emissions, the BB QP limit is 21 dB greater than the NB limit so the single limit approach is not recommended for conducted emissions. However, the study pointed out that an EDP/OE product satisfying the NB limit with a QP measurement also satisfies the broadband limit.
The study recommended that a unit under test should be configured and operated in a manner which tended to maximize its emanation characteristics in a typical application. Power and signal distribution should simulate typical application and usage and at least one module of each type should be operational.
The study also concluded that measured system profiles do not increase over the system profiles synthesized from individual product profiles in a typical system configuration. Therefore, the study concluded that “the worst-case individual product emanation levels of the level of a typical test system can be used to determine the maximum system amplitude level for many possible combinations of products.”
The study also stated that “In the case of test units which functionally interact with other units, either the actual interfacing units or simulators may be used to provide representative operating conditions provided the effects of the simulator can be isolated or identified.”
Also, the study recommended that the measurement site for radiated emissions be a “special environment” to allow valid and repeatable measurements to be made. So the study recommended using a “free-field, non-reflecting electromagnetic environment as closely as possible.” Furthermore, the site should be flat, free of overhead wires, free of nearby reflecting structures, large enough for a thirty-meter measurement and satisfy a space-attenuation of radiated fields. Also, site ambient levels were recommended to be at least 6 dB below the regulatory limit.
The study permitted a “practical site” allowing measurements at a minimum test distance of 3 meters. However, the study found that “results of measurements in such practical sites at varying distances between the equipment being tested and the measurement antenna, have been found to be within +/- 6 dB of those predicted using a 20 dB/decade fall-off relationship between the equipment and the antenna.” The study also said that “Other test sites such as RF semi-anechoic chambers that are properly evaluated to show equivalence to a free-field site may be valid.”
Recommendations for “antenna requirements” included (1) a calibrated dipole (tuned or broadband) for 30 – 1000 MHz and (2) a horn antenna used in the range of 890-1000 MHz. The 50-ohm, 50-uH Line Impedance Stabilization Network described in CISPR Publication 11 is recommended by the CBEMA paper.
Section 11 – Conclusion
The conclusion of the CBEMA report is duplicated in its entirety below.
“This report represents the culmination of many years of effort devoted to an assessment of the potential of EDP/OE products for EMI with authorized communication services. This assessment has shown that the incidence of interference to communications services from EDP/OE has been small and can be kept small in the future by emanation limits specifically developed for EDP/OE products. It is technically and economically important that limits specifically based on the relationships between EDP/OE products and principal receivers be applied. This is necessary because existing and proposed limits not so based, have been found in some cases to either, under-protect communication services, or require unnecessary product emanation suppression, as well as to complicate the interference control process.
Specific environmental as well as emanation source characteristics distinguish commercial EDP/OE from other electrical/electronic equipment. An interference model and emanation limits have been developed specifically for commercial EDP/OE and U.S. communication services; however, the methodology and perhaps the specific limits proposed in this document can be more broadly applied to other types of equipments, and communication services in other countries.
Theoretical analysis supported by empirical data shows that the two types of interference analyzers customarily used in Europe and the USA can both be used to obtain an equivalent degree of interference control. Practical measurement considerations lead CBEMA to recommend the use of interference analyzers complying with the requirements of CISPR 1, 2, and 4; however, other measurement techniques can produce equivalent results.
CBEMA believes that these limits provide more than adequate protection for principal communication services. Both limits and measurement methods have been corroborated by favorable experience, empirical evaluation, and statistical estimates, and are strongly recommended for commercial EDP/OE environments where interference controls are deemed necessary.”
In 1979, two years after the release of the CBEMA paper, the FCC announced its FCC Rules on Computer Emissions. The rules had two dates for implementing the Rules; the first date was October of 1981 when all NEW Electronic Data Processing and Office Equipment had to meet the FCC Rules on Emissions, and the second date was October of 1983 when ALL EDP/OE equipment manufactured after that date had to meet the FCC Rules.
In general, the 1979 FCC Rules imposed on the EDP/OE manufacturers were very close to the recommended CBEMA limits for commercial equipment.
|Daniel D. Hoolihan
is the Founder and Principal of Hoolihan EMC Consulting. He is a Past-President of the EMC Society of the IEEE and is presently serving on the Board of Directors. He is presently an assessor for the NIST NVLAP EMC and Telecom Lab Accreditation program. Also, he is the Vice-Chair of the ANSI ASC C63® committee on EMC.