1958 IRE TRANSACTIONS ON INSTRUMENTATION 297

Principles and Application of RadioInterference Measurements*

W. E. PAKALAt AND R. M. SHOWERS:

INTRODUCTION NEED FOR STANDARDIZATION

TIIHE purpose of this paper is to review the present Standardization in this field is required to establishstatus of radio interference instrumentation, the common terms of reference for industry and for variousproblems associated with its use and standardiza- government agencies including the FCC and the several

tioIn, and to describe the probable areas of future devel- branches of the military services. Persons in all of theseopment. It is timely, for in the very near future the areas are acutely aware of the importance of interfereniceAmerican Standards Association is expected to publish control in their operationis, and of the conisequences oftwo documents in this area as American standards. The failure of adequate control. However, too much conitrolimportanice of stanidardization has been implied in some or too severe requirements are undesirable, since there are

of the previous presentations, but a full appreciation of critical economic factors involved. Everyone agrees thatits importance can only come after a fairly intensive re- a great deal can be done especially if interference controlview of the problems involved. One reason standardiza- is considered and practiced in the development stages.tion is so difficult is the wide variability in the quantity Some of the major factors that are involved in standardi-being measured. What is referred to as radio interference zation will now be considered.can take many forms and appears over the completeradio spectrum ranging from audio frequencies to theINSTRUMENTATIONfar infrared. Lack of standardization has hindered prog- Interference Typesress in this field; however, there is great promise for ad- For convenience, three basic types of interferencevances in this field in the years to come, as a result of have been defined. These are the sine wave, randomASA action. (Gaussian) interference, and impulsive interference.

It is well to point out the relatively large amount Of Practically no source of interference is completely de-effort being made in arriving at international standards, scribed by any one of these waveforms. Almost all sinelargely through the International Special Committee oni waves are modulated either in frequency or amplitudeRadio Interference (CISPR). There are some differ- or combinations of these. Random interference is al-ences in techniques between American and European most always non-Gaussian and impulsive interferencepractices which are difficult to resolve on the basis of may not have a constant repetition rate and almostpresently available technical information. never has a constant amplitude. For purposes of instru-The ASA Standards are C63.2, "American Standard metntation, we further classify interference as narrow-

Specifications for Radio Noise and Field Strength Me- band and broad-band. To distinguish these three typesters, 0.015 to 30 Megacycles/Second," and C63 .4, requires different units of measurement. The significant"Methods of Measurement of Radio Influence Voltage measure for sine-wave interference is in terms of micro-and Radio Influence Field (Radio Noise), 0.015 to 25 volts, but if it is severely modulated a complete descrip-Megacycles/Second, Low Voltage Electric Equipmenit, tion requires some information on the magnitude of thisand Non-Electric Equipment," both sponsored by Sec- modulation. The proper description of the level oftional Committee C63 on Radio Electrical Coordination, broad-band interference is in terms of the spectral intenl-The instrumentation standard reflects accomplish- sity. This could be given in terms of energy, but histor-

ments that have been made in recent years and there is ically it has been given in terms of the microvolt level.Inow a fairly broad understanding of the requiremetnts In the case of random noise, the microvolt level is givenfor obtaining significant measurements of the quantities per square root bandwidth and in the case of impulsiveinvolved and of the limitations in the measurements interference it is given in microvolts per bandwidth.themselves. However, muc.h can be done to bring instru- In measuring the broad-band waveforms, it is essentialment parameters iiito a closer tolerance range, to im- to know the bandwidth of the indicating instrument.prove stability and to simplify calibration techniques. The effective bandwvidths for random and impulsive in-

terference are, in general, different and unless a particu-lar shape factor and phase funlction are adhered to in

* Manuscript received by the PGI, August 20, 1958. Part of the the design of all instruments, the relationship betweenwork reported in this paper was supported by the Navy Dept., ths tw aditsi ohrena osatfo nBureau of Ships, under contract NObsr 72505.thstw addh 1noernarcsalfmoe

t Westinghouse Res. Labs., Pittsburgh, Pa. instrument to another. It is current practice, therefore,t Moore School of Elec. Eng., University of Pennrsylvania,alesintscotr,ospifindiintohe-d

Philadelphia, Pa. 'a es nti onr,t pcf nadto ote6d

298 IRE TRANSACTIONS ON INSTRUMENTATION December

No. of F/imp AF/rn |zF/imp the peak is not a well-defined quantity, and thereforeSest _ it is difficult to establish a standard for its measure-StagesAF/6 db AF/3 db AF/rn ment. On the other hand, peak detectors particularly

1 1.814 1.571 2.0 of the aural or visual kind have been found to be quite42 1 .56 122 1 .473 useful in instrumentation for measuring characteristics10 11.074 1.09 1.42 of various interferences. The other two common types

(a of detectors, average and rms, provide indications not(a) too much different from each other on most types of

1 1 .09 1.09 1 .32 interference. The rms detector is considered signiificant2 1.05 1.04 1 .26 in that it provides a measure directly related to the4 1 .02 1.01 1.21 energy content. It is, however, a more complicated cir-10 0.97 1 .01 1.17 cuit than the average or the quasi-peak, and for this

(b) reason has not found favor with most instrument users.Fig. 1-Relationships between 6-db, effective impulse, and effective The Bureau of Standards in Boulder has been using the

random noise bandwidths. (a) Single-tuned circuits. (b) Double- rms detector extensively in its world-wide program oftuned circuits (critical coupling), atmospheric interference in conniection with the IGY.

It should be noted, however, that atmospherics are ofbandwidth, the effective random noise bandwidth, and a random nature.the effective impulsive noise bandwidth. Fig. 1 shows As noted above, rather than attempt to measurethe relationships between the 6-db, effective impulse, and quantitative interference effects, it appears that theeffective random noise bandwidths for single-tuned and only solution to the "what kind of detector to use"double-tuned interstage coupling circuits [I ], [2]. problem is to try to measure a quantity which can be

directly related to the decrease in the effective informa-tion capacity of the transmission channel due to inter-

In order to obtain quantitative values for the wave- ference. Some progress has been made along these linesforms measured one can choose from a variety of detec- and more is expected. One of the factors is the type oftor circuits. Those used have been described as peak, modulation being used in the signal being interferedquasi-peak, average, and rms detectors, which indi- with. In addition, it is necessary to know certain char-vidually have their proponents and counter-proponents. acteristics of the receivers being used. An alternateHistorically, the quasi-peak circuit was one of the orig- method is the expression of interference effect in termsinal circuits used, probably because it is the most simple of the effects on near perfect receivers receiving nearcircuit-wise of those available, and because it is very perfect types of modulation. Such figures can then besimilar to the type of detector originally used to detect interpreted in particular situations in terms of actualamplitude modulation broadcast signals. In an inter- modulations and receivers used.ference measuring instrument it was recognized that it Some experimental work has been carried out withwas important to provide a response to waveforms both actual and ideal receivers. For the purposes of illus-having a very rapid rate of change, and that this re- tration some of the results obtained to date are shownsponse was an important consideration for standardiza- on Fig. 2 and Fig. 3 [4]. Here the interference is meas-tion. About 1940 listening tests were devised to test the ured in terms of the per cent baud errors when simpleeffectiveness of various charge and discharge time con- binary information is transmitted. With this type ofstants for this detector in providing a reading propor- transmission, the errors are easily identified and cantional to the interference effect of various sources of in- generally be calculated theoretically. It has been foundterference on amplitude modulation broadcast material that good correlation of measured noise voltage intensi-[3 ]. The time constants chosen were 10-msec charge and ties and error probability is obtained when an average600-msec discharge. Since that time, the importance of detector is used. Fig. 2 shows results with a laboratoryinterference to other types of radio communications sig- random noise source, while Fig. 3 shows results withnals has been recognized; indeed, the amplitude modu- tnoise originating in ac and dc commutators. While it islation case is probably of very little significance when too early at the present time to draw conclusions, thissuch items as military defense, FM radio, and television work shows promise for establishing proper measures ofare considered. It now seems more appropriate to try such complex waveforms.to define the effect of interference in terms of the loss ofinformation as defined by information theory. Means of CalibrationThe peak detector attempts to read the true peak Three types of calibrators have been used for stand-

reading of any interference waveform. The philosophy ardizing the response of interference meters. These typesunderlying this detector is that if the peak is suppressed, correspond to the basic interference types, mainly, sineit is unlikely that the interference will be troublesome. wave, random, and impulse. The sine wave type is theThe difficulty with this philosophy, however, is that for most complicated of these circuit-wise since it has to besome interferences, particularly random interference, tuned. The other two types are, of course, broad-band,

1958 Pakala and Showers: Principles and Application of Radio Interference Measurements 299

50 F only provides convenience in use, enabling one to read arangeof two decades without switching attenuator posi-

40C POINTS' tions, but also provides for a constant indicator accuracyo EXPERIMENTAL P0.1NTS 0- THEORETICAL CURVE over the scale. In view of these problems, it is not sur-

30 1 _ __ prising to find that an interference measuring set costsI I I / considerably more than a good receiver in the same fre-

co / quency range.

0

0 / As indicated previously it has not been possible to es-.05 IAVERAGE VOLTAGE RAT (NOIS1 SIGNAL tablish a single type of band-pass characteristic to be

used in these instruments. At the present time, commer-Fig. 2-Error probability on binary communication system cial instruments have effective bandwidths varying over

with a laboratory random noise generator.the range of three to one, and show quite serious varia-

40 tions of the ratio of effective impulse bandwidth to0 A.C. COMMUTATOR MOTOR effective random noise bandwidth. There seems to bex D.C. COMMUTATOR MACHINE 0

30 insufficient information available to determine whatwould be the most desirable shape to use. Manufac-turers are reluctant to meet any rigid specification re-

qm 20m / quirements because of the additional cost which may be

involved, and the users are reluctant to share such aI/ burden, also because they do not have adequate infor-

mation available on the important types of interferenceo _ _ . so they can make a decision as to what is the most de-.05 .2 .5 21 5 tcncly tematm,~ seiiain

AVERAGE VOLTAGE RATIO (NOISE/SIGNAL) sirable technically. In the meantime, ASA specificationsFig. 3-Error probability on binary communication system previously referred to will contain requirements to the

for interference originating in commutators. effect that all manufacturers shall state the 6-db, therandom noise, and the impulse bandwidths for their

and when used enable standardization of the product of instruments.gain and bandwidth or gain and square root bandwidth.Satisfactory random noise sources have not been pro- Detector Circuitsduced. The outputs available are too small and power Present specifications permit the use of any one or allrequirements are too large. The impulse calibrator has of four detector circuits-peak, quasi-peak, rms, andbeen more satisfactorily developed in recent years. It average. They do state, however, the characteristics ofprovides a reasonably large output, relatively small size, such circuits if they are used, and it is presumed thatand perhaps the best frequency range, at least for in- the users of the instruments will tell the manufacturersternally contained calibrators. Further developments in what circuits they require.this type of calibrator are needed to increase the fre-quency range, to provide an output clear of secondary Antennasemissions, and to reduce costs. The antenna situation is one of the weakest in the

whole instrumentation problem, as contrasted to thesituation in measuring field strength, where the meas-

The usual type of circuit used in measuring instru- urement of the magnetic field is adequate, partly as aments is based on the superheterodyne. In addition to result of the fact that the measurements are made in theconsideration of effective bandwidths and detector cir- radiation field of a distant antenna. For interference,cuits, the designer is confronted with problems associ- requirements exist for the measurement of fields in theated with quantitative measurements such as the in- near field region. The most sensitive type of antenna atcorporation of an attenuator to enable voltages from a frequencies below 30 mc has been found to be the rodmicrovolt to about 10 volts to be measured, obtaining or electric type. As compared with the loop type, its re-a relatively constant input impedance over a wide fre- sponse is very sensitive to ground plane characteristicsquency range for all attenuator positions, obtaining and its impedance is high. This requires the antenna tosensitivity equal to that of any receiver operating in be placed close to the instrument, and it has becomethat frequency range, and linearity of indication with common practice at frequencies below 30 mc to mountinput level to all kinds of waveforms. In addition, most it directly on the instrument case. The uncertainty ofinstruments in this country use automatic gain control the ground plane under these conditions, and the effec-to provide a logarithmic scale. The logarithmic scale not tive distortion of the field provided by the presence of

300 IRE TRANSACTIONS ON INSTRUMENTATION December

Characteristic ~~~[5] 6] [7]Characteristic ASA CISPR U.S.S.R.

4-8 kc (nominal) +1 percentBandwidth (6 db) Variations permitted, give effective impulse 9 kc ± 10 per cent 9 kc

and RN bandwidths -30 per cent

Detectors Peak Quasi-peak Qtuasi-peakQuasi-peak (Peak, rms and average may berms included)Average

QP Timne ConstantsCharge time (msec) 1 1 10Discharge timi-e (msec) 600 160 600

Indicating meterTime constant (nmsec) 200-350 160 200-400Permissible overshoot 1-6 per cent critically damped 2-5 per cent

Antenniia i meter rod 1 meter rod 1 meter rod

Fig. 4-Comparison of radio interference inistrument specificationis.

the instrument leaves a great deal of room for improve- should be noted that generally specifications in foreignment in the accuracy of the results obtainable. Means countries have been simpler than those in this countryshould be established for making measurements in the until very recently. Provisions are made for only onenear field of the electric components with accuracy in type of detector, and generally a linear scale calibrationthe order of 1 db. has been used. Another difference is that the impedance

across which measurements are made is 150 ohms inSummary of Status of Standardization Programs Europe as compared to 50 ohms in this country. OtherThere are two organizations in this country which are differences are illustrated in Fig. 4 which shows a com-

leading in standardization work: one is the American parison of proposed American Standards, those ofStandards Association, and the other is the BuLreau of CISPR, and those formulated in the Soviet Union. ItStandards. Work by the Bureau of Standards on atmos- appears that it will be several years before true inter-pherics has already been mentioned. The techniques national standardization is accomplished in any fre-being used, which briefly consist of the measurement of quency range. However, because of a great amount ofthe rms value averaged over a period of several minutes, exchange of information which is currently going on, therepresents a culminationi of effort of a truly international future looks bright.nature on this particular problem.The work of the American Standards Association is APPLICATIONS

continuing in its area of measurement with an active Radiation tests in manufacturing and test areas re-program for extending the frequency range beyond quire shielded enclosures to obtain the low ambientthose of present standards. levels necessary to determine if equipment meets radia-Two groups have been active in international stand- tion limits.

ardization. The International Electrotechnical Com- Shielded enclosures are constructed in several differ-missioni, through Subcommittee 12-1, has been dealing ent ways, such as: 1) solid copper single-wall and double-with the problem of the measurement of spurious radia- wall; 2) copper screen single-wall and double-wall; 3)tion from AM, FM, and TV receivers. A new document solid-steel single-wall; 4) double-wall steel and copper.will shortly be issued as a recommended standard. The All these enclosures, when correctly installed, haveUnited States (IRE) has similar standards and has ob- proven satisfactory for measurements in the frequencyjected to the recommended standard. The major point range of 0.014 to 1000 mc in and near manufacturingof issue is the distanice for making radiation measure- areas. The choice of wall material depends at times onmenits. Most of these measurements are made at a dis- the environment and its effect on life of enclosure, as-tance of 100 feet in an open area, whereas this standard sembly at location, ease of moving to new location, mag-calls for a measurement at three meters. The United netic field attenuation, ventilation, size of enclosure de-States is reluctant to go along with this proposal until sired, and maintenance. It is important that the filtersmore quanltitative measurements on its effectiveness are in the power-supply leads be specifically designed foravailable. shielded enclosures and that they be properly installed.The CISPR previously mentioned is more directly Afa

concerned with measurement of more genleral types Of ASaurements on Low-Voltage Apparatusradio interference. Through it, meetinlgs of international Measurements on low-voltage (below 600 volts) ap-complexion are being held several times a year, and paratus are considered separately from high-voltage ap-there is much exchange of technical information. It paratus due to differences in measurement methods.

1958 Pakala and Showvers: Principles and Application of Radio Interference Measurements 501

To Test 7 From Power II_I _Sample - 0 Supply or Lood 36 it

Olp.fd-~~ L 5 p. Henries Oh K(+20%)0.1ifd.~ (when mou ted TI Ohm 48__

1000 Ohms + . in shielding E Designenclosure) Ca-7cesifdcj

040

/4 t----____4; a @ { |/ /1 ~~~~~~~~~Min.Limit||Ground Plane Ground Lead. Ground Plane g32

-oShielding Enclosure

50 Ohm Coax 50 Ohm Termination E 24

lOiip.Lfds _____RodioalNoise __Meter e

Fig. 5-Line impedance stabilization network used forconducted measurements on low-voltage devices. O 1 10

Frequency, Megacycles/Second

The conducted-voltage measurements at terminals are Fig. 6-Input impedance characteristic of linemade using the radio-interference instrument as a volt- impedance stabilization network.

meter, by connection through a capacitor to the deviceterminal. A line-impedance stabilization network is con-nected in each line lead so that a definite radio-frequency Rod Antenna

loading is maintained.~~~~~~~~~~Reel Test Fixture _ReversinglOading iS malintained. Device / 50 Ohm /,-Coax Switch

Fig. 5 shows the circuit of the line impedance stabili- Test TerninatonOzation network and Fig. 6 shows the input impedance 24"g Dis1a co 10Coaa fdscharacteristic of this network including the impedance Meter Cor MeterMRdaedFltrrd Mer(onuterlimits permitted over the frequency range of 0.15 to 25 |(Rasdied | ToFiate redPower (Conductedmc. The impedance depends on the radio noise measure- Ground Plane K Line Impedancement frequency up to a few megacycles and at higher Stabilization Networksfrequencies it is substantially constant. The character- Fig. 7-Radiated and conducted measurement-singleistic shown is used because of load current and power phase ac or two-wire dc.frequency limitations of a constant impedance network.The measurement of radiated or coniducted radio Measurement of Interference Output of Television Re-

noise from a device may be required. Fig. 7 shows a test ceivers in the Range of 300 to 10,000 KC, 1954. IREsample mounted on a standard insulating reel type test Standard, 541RE17.S1. A supplement has been pre-fixture with test sample line cord wrapped around the pared for this standard.fixture. In this figure a radio noise meter with a vertical Considerable instrumentation and space are requiredantenna is shown three feet away for the radiation for these tests and studies are being made to find out iftest with connection of meter to stabilizing network for the space requirements can be reduced. Because of am-measuring conducted radio nioise voltage from each test bient disturbances in most manufacturing and electricalsample lead to ground. test areas, it is desirable but not always necessary toOn large devices the tests are to be made using reel make conducted-voltage tests on machinery in a

test fixture for supporting the line cord; however, the shielded enclosure. If the power-line leads are filtered ondevice is restinig on the ground plane, on 3-inch thick the load or power-supply side of the stabilization net-insulating plate. If conducted measurements are made works, conducted tests often can be made eveni to thein, the field onl installed devices a 50-ohm probe with a lowest limits specified.suitable coupling capacitor to power line is used to con- A shielded enclosure or an open space with low ambi-nect meter to the line or to the device terminal. The ent level is almost a necessity for radiated tests. Ma-methods described above can be found in the proposed chine test areas in factories are unsuitable for militaryASA Standard C63.4. specification radiated tests. For machines that cannot

Conducted measuremnents in future ASA Standards be transferred from shop test area to a shielded enclo-will probably be extended up to frequencies where it is sure, a portable shielded enclosure is assembled aroundknqown that conducted measurements are important in the machine. Open space locations 200 feet from build-the control of radio interference. ings and away from lines, fences, etc., are suitable for

For some low-voltage devices such as TV receivers, radiation tests at short distances, up to 100 feet. How-special measurement procedures have been developed ever, such locations have disadvantages due to weatherby the IRE. Both radiated and conducted are measured and to time required to set up and transport equipment.at all frequencies of interest. These methods are de- Therefore, indoor facilities and test techniques equiva-scribed in IRE Standards Onl Receivers: Methods of lent to an open space are desirable and often necessary.

502 IRE TRANSACTIONS ON INSTRUMENTATION December

High Volt ge RFCBUS-1 ~~High Voltage Bus

TestTransformer Device Under Test

Coupling 600w RFCCac000ors

].Rodio

insulation@1 MNose

Fig. 8-Test setup for high-voltage devices.

The military services have provided much informa-tioni in the development of techniques described. They)have made studies of transmitters, receivers, generators,instrumenitation and measurement technuiques, installa-tions, and developed specifications and limits.

Measurements on IHligh- Voltage Apparatus Fig. 9-Radio inflticuece test at high voltage on a conductor.Most tests on high-voltage apparatus use the test

circuit specified in "Methods of Measuring RadioNoise," 1940, EEI Publication No. 69, NEMA Publi-cation No. 107, RMA (now EIA) Eng. Bulletin No.32. The circuit used is shown by Fig. 8; note the locationof the 600-ohm resistance across which the radio-influence voltage is measured. This method is a coIn-ducted measurement for a limited frequency range of0.54 to 1.6 mc. Most measurements are made at 1 mc,since the NEMA standards for radio-influence limitsspecify this frequency. This method, although not asdirect as a radiation test at the installation, is useful inthe design of high-voltage apparatus to specified limits.It is quite sensitive and can be used to obtain, in somecases, the corona threshold voltage unider low ambienitnoise conditions. This NEMA method has been used upto 400 kv rms to ground in laboratory test setups, andeven higher test voltages are practical. All conductors,guard rings, insulators, anid capacitors used to build this Fig 10-Radio iiiflttice test at high voltage on a disconiiect switch.circuit should be free of corona at all test voltages. Con-ductors and guard rings used in the test setup requirefrequent polishing and cleaning to maintain them in a Acculracycorona-free state. ASA Specification C63.2 for Radio Noise and FieldAlthough tests have been made in high-voltage lab- Strength Meters does not give any single percentage ac-

oratories with very little wall shielding, experience indi- curacy figure. Instead the pulse response at 20 pps rela-cates that a shielded enclosure is definitely desirable for tive to 100 pps is given along with upper and lowerlaboratory tests and new high-voltage laboratories have limits. The maximum difference between upper andbeen shielded and the power supplies have been filtered. lower limits is 1.5 db. The specification for InternationalFigs. 9 and 10 are photographs of radio-influence tests Special Committee on Radio Interferenice (CISPR) ofat high voltages. the IEC specifies accuracy as follows:

Radiation Measurements Required by the FCC 1) The accuracy of measurement of sine-wave volt-

FCC regulations have been developed to limit radia- ages shall not be worse than ± 2 db.tioni from RF heaters used for induction heatinig of 2) When connected to a suitable antenna the accu-

metalsOII d dielectric heatiiig of plastics; radiation racy of measurement of the strength of a uniformmetals and dielectric heating of plastics; radiation sine-watve field shall be not worse than ±+3 db.measurements are made on heaters manufactured, andon heaters installed for various manufacturing processes. For regularly repeated pulses the response to pulses ofHeaters are also inspected periodically and a log kept 0.316 ,uv seconds having a uniform spectrum up to atof these inspections. Radiation limits are included for least 30 mc, repeated at a frequency of 100 cps, is to beother types of devices also as defined in Parts 15 and 18 equal to the response to an unmodulated sine-waveof the FCC Rules. signal of value 2000 ,uv from a signal generator having

1958 Pakala and Showers: Principles and Application of Radio Interference Measurements 505

70

600

2 50 100

-Oo qspQuasi Peak- - m

40~~ ~~~~~~~~~

C~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-

20

510 30 00 300 1000 -#- Ic -- -Pulse FRepetition Pate -Pulses/Seco'nd

Fig. 1 I-Pulse response comparison of three meters Con quasi-peak at 1.25 mc.

the same output impedance as the pulse generator. Atolerance of + 1.5 db is allowed on the above voltagelevels.

For variation with repetition frequency of pulses a -l0 l0 - - lcurve is given and the receiver's response should follow .1

-

10this curve within limits from + 1 db to + 2 db depending Frequency - Megacycleson the repetition rate of the pulses. Fig. 12-Comparison of three NM-20B meters. Meters read simul-The pulse response of three radio noise meters made taneously under double-circuit vertical configuration transmissionline.

by one American manufacturer is shown in Fig. 11.These three meters were also compared on line noise BIBLIOGRAPHYusing vertical antenina. Meters were read by different 1] R. M. Showers and A. Eckersley, "Research investigations ofoperators and compared as shown by Fig. 12. The max- interference measuring instruments," Proc. Conf. Radio Inter-imuiim differences between mneters was, in this case, 3 db. ference Radiation, Armour Inst. Tech., Chicago, Ill. 1954.

V2]Y. Peless, "Response of cascaded double-tuned circuits," to bepublished in Electronic and Radio Engineer.CONCLUSION [3] C. M. Burrill, "An evaluation of radio-noise-meter performancein terms of listeningexperience," PROC. IRE,vol. 30, pp. 473-478;

The great expansion in the use of the frequenicy spec- October, 1942.trum has increased the complexity and number of inter- [4] J. T. Suss and F. Haber, "Error probability in a simulated binarycommunication system," to be published.ference problems and actions of the FCC and the mili- [5] "Draft of American Standard Specification for Radio Noise andtary services. Field Strength Meters 0.015 to 30 Megacycles/Second C63.2;"

November, 1957.The control and measurement of the initerference re- [61 "Draft Specifications for CISPR Radio Interference Measuring

quires special techniques, which are not always appre- Apparatus for the Frequency Range 0.15 mc/s to 30 mc/s;"January, 1958.ciated nor always fully understood. Although advances [7] L. G. Robbins (University of Alabama), "Standards of Permissi-have been made in the radio-interference field, much ble Limits of Industrial Radio Interference." Translated from

more rsearc and oint ctionby tehnica grous in- Russian; 1956.more research and Joint action by technical groups in- [81 "American Standard on Methods of Measurement of Radio In-volved is needed to simplify instrumentation, test pro- fluence Voltage and Radio Influence Field (Radio Noise), 0.015

to 25 mc/s Low Voltage Electric Equipment and Non-Electricedures, anid improve specifications. Equipment-ASA c63.4", November, 1957.