Bell Laboratories Measurin G Sky'ss 'S Electrical Noise `4 Surface -To -Air Data Communications the Art of Counting Calls a Statistical Ammeter

L SEP '8A IBAY{ Gy August 1961 Electronic Memory Devices (4B aQ- Bell Laboratories Measurin g Sky'sS 's Electrical Noise `4 Surface -to -Air Data Communications The Art of Counting Calls A Statistical Ammeter

www.americanradiohistory.com F. J. Singer, Chairman W. M. Bacon J. A. Burton Editorial Board J. W. Fitzwilliam E. T. Mottram R. J. Nossaman W. E. Reichle

H. W. Mattson, Editor A. G. Tressler, Associate Editor Editorial Staff J. N. Kessler, Assistant Editor, Murray Hill M. W. Nabut, Assistant Editor T. N. Pope, Circulation Manager

THE BELL LABORATORIES RECORD is published monthly by Bell Telephone Labora- tories, Incorporated, 463 West Street, New York 14, N. Y., J. B. FIsK, President; K. PRINCE, Secretary; and T. J. MONTIGEL, Treasurer. Sub- scription : $2.00 per year; Foreign, $2.95 per year. Checks should be made payable to Bell Laboratories Record and addressed to the Cir- culation Manager. Printed in U. S. A. © Bell Telephone Laboratories, Incorporated, 1961.

www.americanradiohistory.com Bell Laboratories RECORD

Volume 39 Number 8 August 1961

Contents

PAGE

270 Electronic Memory Devices D. H. Looney

276 Measuring the Sky's Electrical Noise D. C. Hogg and H. E. D. Scovil

280 Portable Ruby Optical Maser Demonstrated

281 A New Surface -to -Air Data Communication System J. B. Bishop

285 The Art of Counting Calls W. B. Callaway

289 A Statistical Ammeter J. W. Osmun

292 Ferrite Isolator: New Kind of Pad for TJ Radio -Relay A. F. Pomeroy

This ferrite sheet, with its printed wiring contacts, is part of a memory Cover module -one of the many types of memory devices discussed in a review article which begins on page 270.

www.americanradiohistory.com Latest development in twistor memories for large wImt, demonstrated by U. F. (; ?Iola (left) and J.A. capacity storage is this card -changeable arrange- Ruff, Solid State Device Derel- )pment Department.

270 Bell Laboratories Record

www.americanradiohistory.com Automatic communications systems, with their emphasis on speed and reliability, need fast and dependable memory devices. ¡"or this purpose, communications technology/ has presented the systems designer with several interesting choices.

D. H. Looney

Electronic Memory Devices

Telephone switching systems, and other com- that identifies the physical location of the incom- munications systems of many types, are now being line on a particular frame of equipment. This ing developed to operate at extremely high speeds, number will exist unchanged until the customer measured in millionths of a second. The reason moves or requests a different class of service, or for this emphasis on speed is partly a desire to until the offices revises this equipment number perform communications services faster, but it is for some reason. also a desire to process a given amount of infor- Another form of information that may be stored mation with less equipment. Thus, if the prices for a considerable period is an "instruction." An of high -speed circuits can be held to reasonable ordered set of instructions is a "program" which levels, significant savings may result. prescribes the decisions and sequence of opera- "Memory," or storage of information, is only tions necessary to accomplish the designed pur- one part of a communications switching system, pose of the system-processing telephone calls. but it is a very important part. It ranges from Present -day telephone offices are programmed storing an item of data for only a very short entirely by wiring patterns interconnecting period -maybe only a fraction of a second -to switches such as relay contacts. Programs for storing another item for perhaps days or months. future telephone offices will be stored in electronic For example, when a customer dials a telephone memories as lists of instructions and changed number, memory circuits in the local central office whenever new features or services are added. may be required to store all or parts of this num- These concepts are important to an appreciation ber for only a few seconds. On the other hand, of the many new methods of high- speed, electronic other data may be stored indefinitely. The line memory. They explain, for example, the classifi- coming into the office from a telephone, for in- cation into "temporary" and "permanent" mem- stance, has an "equipment number " -a number ories, and they introduce the important role that

August 1961 271

www.americanradiohistory.com memory plays in directing a stored program electronic switching system. In this discussion, it would be impossible to include every new device and idea being considered for memory applications. Also, in such a rapidly changing field, it is difficult to gain enough perspective to separate the significant developments from the many interesting proposals that may never become economically practicable. The few examples here, however, should give some of the flavor of contemporary work on memory systems, and should emphasize the significance of modern electronics for future communications. Aside from the general distinction between temporary and permanent memories, storage systems can be categorized in many ways. One of the most basic is to think of them according to the phenomena responsible for the storage. We can imagine almost any discrete effect which exists in two states -for example, a liquid solution can be either electrically conducting or nonconduct- on these photo- ing. As long as the effect can swing or be Bits of information are recorded tiny clear or opaque dots. "switched" from one state to another, it can be graphic plates as used to design a memory. Two effects used in electronic memories now altered. In this manner, the spot stores a small under field trial in the Bell System are the pres- quantity of electrostatic charge. Its presence or ence or absence of an electrostatic charge on a absence specifies one binary digit, or "bit" of sheet of an insulating material, and the presence information. or absence of an exposed area on a photographic In barrier -grid tubes now being studied, a film. The electrostatic phenomenon is the basis for typical storge array consists of a square pattern the design of a barrier -grid temporary memory, of some 16,284 spots (128 by 128) . Each spot is and the exposed film is the basis of the flying -spot only about 5 thousandths of an inch in diameter. store or permanent memory -both used in the Because of such small dimensions, the electron electronic central office currently under field trial beam must be positioned and deflected very pre- at Morris, Illinois (RECORD, December, 1960). cisely. After a charge has been stored on a spot, it Barrier -Grid Store can be removed ( "read out ") and a new charge The barrier -grid store is a system having a inserted ( "written in ") in a cycle time of 21/2 the vacuum tube in which a beam of electrons is di- millionths of a second. The beam can detect rected to the surface of a sheet of mica. Im- presence of charges ( "interrogate ") on only one mediately in front of this sheet is the "barrier area at a time, and can thus read out only one bit -grid tube is grid " fine screen or mesh of wires that at a time. Consequently, the barrier -a bits controls the secondary electrons falling back on particularly useful in systems which handle tubes are the mica sheet when the state of the memory is serially. If a number of barrier -grid changed. The homogeneous mica surface is di- operated in parallel, however, the output is a read vided into a large number of discrete storage "word" of several bits of information, all areas or "spots." When the beam impinges on out simultaneously. a particular area, the average number of elec- In the permanent memory of the experimental trons arriving at the surface would ordinarily Morris Central Office, the data bits appear as equal the number of secondary electrons leaving. transparent or opaque areas on a photographic However, a metal plate is mounted in back of the film. Like the mica sheet of the barrier -grid tube, mica sheet, and when the voltage on this backing a single section of film is divided into a square plate is properly adjusted, the number of elec- array consisting of a large number of storage trons which stick to the mica can be temporarily areas. In this case, however, the information is

Record 272 Bell Laboratories

www.americanradiohistory.com FORWARD LOSS REVERSE LOSS RETURN LOSS

40 1.2

35 1.0

30 0.8

25 0.6

20 0.4 11.8 11.0 11.4 11.8 10.6 11.0 11.4 11.8 10.6 11.0 11.4 FREQUENCY IN KIIOMEGACYCLES PER SECOND changes Loss curves for a typical lA Isolator. The curves most of the frequency band. Temperature the performance. reveal a "front -to- back" ratio of 2,690,000:1 over do not appreciably degrade

A permanent mag- zinc alloy. These covers support the pole pieces or by ac or dc magnetic fields. it is partially demag- and keep dust from gathering on the magnets. net is said to be stabilized if fully magnetized. The hys- The pole pieces, made from cold -rolled steel, are netized after being a stabilized permanent aligned with the ferrite bars to achieve the de- teresis characteristic of as that of a sired magnetic field configuration. magnet has the same general shape The operation of a field displacement isolator fully magnetized magnet. stabilization is previously (RECORD, November, A simple way of thinking about has been explained magnet 1957). In this rapidly expanding art, theoretical as follows. To magnetize a permanent is applied. This considerations are subject to continuing evolu- fully, a very strong magnetic field overcoming the resist - tion as experimental observations add to the fund field can be thought of as most stubborn single do- of knowledge. Loss mechanisms in field displace- ance-to- rotation of the finally all aligned. Now, if a ment isolators may now be described as follows : mains until they are is applied, In the illustration on page 292 note that each smaller field in the reverse direction how they were ferrite bar is in the field of one permanent mag- some of the domains will remember and will rotate net. Since the direction of circular polarization of orientated before magnetizing, As stronger TE10 mode is opposite in the two sides of a rec- back to their original orientations. a more do- tangular waveguide, the magnets of a 1A Isolator and stronger reverse fields are applied, back. In the stabilized condi- oppositely ; that is, a north pole of mains will rotate are assembled back should be opposite a south pole of the tion, no additional single domains will rotate one magnet than that other magnet. In one direction of transmission, for any reverse field of less strength waves react vigorously with the magnetic used to stabilize the magnet. the of man- fields of the electron spins in the magnetized fer- To check the operating characteristics interactions form electric "walls" on ufactured assemblies, ten Western Electric iso- rites. These for inside faces of the bars, thus forming a rela- lators were measured in Bell Laboratories the of a tively lossless duct between the bars, through electrical performance. The characteristics pass practically unimpeded. typical isolator are shown in the graph above. which the waves for In the opposite direction of transmission the This graph shows results of measurements at a room tem- waves may be thought of as being "scattered" many frequencies across the band F and into the ferrite by the pattern of chrome coating, perature of 75 degrees F ; but at 40 degrees made at much as light rays are deflected by a venetian at 140 degrees F, measurements were blind. These scattered waves are dissipated in the band -edge frequencies only. The results indicate ferrite bar, causing high reverse loss. In addition, that the isolator will be satisfactory over this there is some dissipation in the chrome coating. temperature range. precise Since the performance of an isolator depends The lA Isolator is only one of the many fine upon the strength of the magnetic fields in the components that contribute to the perform- ferrite bars, the permanent magnets in the new ance of the TJ Radio System. Because it is a isolator are stabilized to prevent changes in oper- marked improvement over conventional "pads" of ating characteristics because of magnet aging. resistive elements, it represents an important This aging may be caused by mechanical shock, step forward in microwave communication.

Record 294 Bell Laboratories

www.americanradiohistory.com transmission, there is little or no scattering, hence, almost no absorption. The effectiveness of an isolator may be described by stating its forward loss, reverse loss and return loss. An ideal isolator, for example, will cause no attenuation in the transmitting direction ( zero forward loss) and complete attenuation in the reverse direction (infinite reverse loss ) . Further, of all power incident upon the isolator, none would be reflected by the isolator itself (infinite return loss). In the actual TJ design, the 1A Isolator has (within the 10.7 to 11.7 kmc band) a forward loss of about 0.7 db, a reverse loss of about 65 db, and a return loss of about 32 db. This means that of the available signal power, about 85 per cent is transmitted, whereas only about 0.00003 per cent of the power reflected in the waveguide run or at the antenna reaches the klystron. A return loss of 32 db means that only about 0.06 per cent of all the power which reaches the isolator from either direction is reflected from the device. The com- A. R. Johnson, show', iog (!d/,i/ a complete lA Iso- plete TJ isolator (see drawing on page 292), con- lator. Alternating ein rent is used to stabilize the sists of two ferrite bars, a section of waveguide magnets to the operating values so that the mag- in which these are mounted, and two permanent nets will remain permanent during years of use. magnets with their associated covers and pole pieces. Each ferrite bar is of the nickel -zinc type with opposite direction. By using ferrites, ceramic - a layer of chrome evaporated in a pattern onto like magnetic materials, losses can be decreased in one broad face. This pattern of chrome scatters the direction of transmission, and at the same the microwave and causes the high reverse loss. time the loss can be increased in the opposite The inside dimensions of the waveguide section direction. The device achieving these unusual are 0.9 by 0.4 inch, held to within 0.0015 -inch. properties is a combination of ferrite and a section This dimensional accuracy is required in such a of waveguide -the combination being called an precise device as the lA Isolator. "isolator." Such an isolator is used between the The half -cylindrical permanent magnets -which klystron and the antenna waveguide run in the supply a steady -state magnetic field fabri- new -are TJ radio -relay system. The complete isolator cated from Alnico V because of its superior prop- is shown in the picture above. erties. The material is cast into final shape be- This type of isolating action can be obtained in cause it is very hard. Critical surfaces are three different ways : first by Faraday rotation ground smooth. Its magnetic properties are ob- (RECORD, October, 1955) ; second by ferromag- tained by heat -treating the finished part under a netic resonance; and third by field displacement magnetic field. In effect, the heat treatment di- (RECORD, November, 1957). vides the alloy into plates of precipitate arranged The third method -field displacement -is em- in a matrix of rods. The spacing between rows of ployed in the TJ system. Here, a transversely - rods is about 200 angstroms. According to mag- magnetized, thin rectangular ferrite slab is lo- netic theory, this small spacing makes it seem cated so that its broad face is parallel to the probable that the magnetization is accomplished, narrow walls of a rectangular waveguide assem- in a finished permanent magnet, by rotation of bly. The position of the slab is critical. It must single magnetic domains, rather than by move- be in a plane at which the wave is circularly ment of domain walls. The magnetic field applied polarized. A more or less conducting film, in a during heat treatment assures that the rods form particular pattern, is placed adjacent to one broad in the most favorable orientations for making the face. The film causes the wave -for one direction best permanent magnet. In general, the axes of of transmission -to be scattered into the ferrite the rods are all parallel. where it is absorbed. In the other direction of The magnets are protected by covers made of a

August 1961 293

www.americanradiohistory.com Pads -networks of resistors -are often inserted into telephone transmission circuits to "smooth" the circuit impedance characteristics. To reduce the insertion loss, get increase the "smoothing" action, Bell Laboratories has perfected ...

A. F. Pomeroy

The Ferrite Isolator: A New Kind Of Pad for TJ Radio-Relay

of Modern communications depend heavily on larities in transmission systems at frequencies elements called microwave radio transmission. In the Bell System, 70 me or less, networks of resistive the transmis- more than half of all long- distance telephone "pads" are commonly inserted into a the are over microwave radio -relay routes. sion circuits. The greater the loss of pad, circuits but In addition, as the microwave art advances, more effective it is in reducing such effects, loss radio -relay systems are becoming more economi- the less desirable it is from a transmission be made up by cally attractive over the shorter distances. For standpoint, since this loss must this reason, Bell Laboratories developed the TJ increased amplifier gain. : in- radio -relay system "short- haul" system Pads are inherently two -way devices a pad -a - of trans- which operates in the 10.7 to 11.7 kmc band troducing 3 db of loss in one direction the same loss in the (RECORD, April, 1959). mission will also introduce Its source of signal power is the klystron. This could be directly connected to the antenna wave - guide run if there were no reflections from the antenna to cause a signal distortion. Signal dis- tortions do occur, however. To minimize them within a radio -frequency channel-and at the same time minimize crosstalk between telephone channels-the frequency- versus -repeller voltage characteristic of the klystron should be linear. ( The klystron's repeller electrode is modulated by a baseband signal that may comprise several hun- dred telephone channels or a television signal.) Unfortunately, this characteristic depends on the impedance of the load to which the klystron is connected; that is, the waveguide run and antenna. The distorting effect of the antenna system is, in effect, proportional to the reflection coeffi- cient of the antenna and the length of waveguide The complete isolator consists of two ferrite bars, run attached to it. a section of waveguide in which these are mount- To lessen the effects of such impedance irregu- ed, and two magnets with pole pieces and covers.

Bell Laboratories Record 292

www.americanradiohistory.com that the unit is light and small. The circuit can

be powered from a 48 -volt central -office battery. BLOCKING Two type -D dry cells supply the bias voltage for OSCILLATOR the electronic gates. Each of the meter's three counters has a BINARY limit of 999,999 counts. If they count steadily II DIVIDER at five counts per second, they can operate for STEADY about 55 hours. However, most test runs will COUNTER take an hour or less. The counters, which are 20'1., LOAD i similar to auto mileage indicators, can be manu- CIRCUIT ally reset to zero. The controls for adjusting RANGE the sampling frequency and for setting the SELECTOR threshold level of the gates on one are side of AMPLIFIER the meter and all the inputs, along with the 120ti milliammeter and the control for the dummy load, RECTI FIER are at the rear. The range- selector switch and various power switches are at the front. óc When the meter is correctly calibrated, the counters 75 -0 start counting at a current level that GATE ICOUNTER may be as much as 5 per cent below the indicated r value. As a result, the recorded count will tend to be slightly high. Tests show that the per- 100% 100% OR 50 %I centage of recorded time that the load current GATE COUNTER exceeds the preselected current levels is within 5 per cent of the correct value. This error is always on the high side so the amount of time 50% that the load current exceeds the selected values GATE is slightly less than the value indicated by the meter. In this way, the premeditated error in Diagram of the statistical ammeter. The three the meter always produces conservative data. gates operate when the ringing current exceeds The statistical ammeter cannot be used to .50, 75 or 100 per cent of the generator's current measure the ringing drain in small PBX ringing rating. The counter records duration of current. plants where the load current is no more than 50 to 100 milliamperes. Laboratories engineers time the busy -hour load exceeded or equaled one feel that it is not usually necessary to measure the ampere, and 60 per cent of the time the current ringing drain in the large 6- ampere ringing plants. exceeded or equaled %4 ampere. If the need arises, however, this meter can be Now, if another busy -hour run is made with modified to extend its range in both directions. the meter still on the one -ampere range and the There is a variation in the 20 -cps waveform 75 per cent and 50 per cent counter in use, the with different types of ringing generators. It results may be something like this: Sixty -two is necessary, therefore, to check the meter cali- per cent of the time the ringing current exceeds bration each time it is used on a different code or equals s/4 ampere and 96 per cent of the of ringing machine. Then the meter may be con- time the current exceeds or equals 1/2 ampere. nected in the ringing load circuit and used for Thus, by selecting suitable ranges, the maximum two or three short trial runs to determine the current can be determined and the load current approximate range. Once the correct range is can be measured as closely as desired. The data selected, a full busy -hour run may be made. At obtained are sufficient to determine how heavily the end of the run, the results are obtained sim- the ringing machine is being loaded and whether ply by taking the ratios of the 50 per cent, it is necessary to increase the capacity of the 75 per cent, and 100 per cent gated count and the ringing power plant. count of the reference counter. The new meter is a simple and accurate method For example, assume that the ringing plant of measuring current where the load fluctuates is rated at one ampere. The meter is set on rapidly. This meter makes it possible to design the one -ampere range and the 100 per cent and efficient telephone ringing plants and thus mini- 75 per cent counters are used during the run. mize service failures and reduce maintenance The results may show that 10 per cent of the caused by overloaded ringing facilities.

August 1961 291

www.americanradiohistory.com cause of low-voltage. These low- voltage condi- ing current five times a second. If the current tions are often caused by overloads. But they equals or exceeds certain preselected levels dur- are relatively minor problems. If an area grows ing the run, an electromechanical counter records at a rapid rate, the power rating of the ringing these levels as "interval counts." The preselected generators should be stepped up proportionately. levels, which are based on the size of the gen- The best way to determine the amount of erators, are chosen with a range -selector switch. ringing current passing through a central office To find out the proportion of time that the is not to estimate it, but to measure it. This ringing current exceeds the preselected levels, avoids a considerable amount of calculation and the interval count is compared to the count from is inherently more accurate. By measuring the the reference counter. The results are expressed ringing current during the busy hours, an oper- as the proportion of time that the ringing cur- ating company knows without question whether rent equals or exceeds 50 per cent. 75 per cent, the generators at a specific central office are and 100 per cent (or any other preselected per- supplying enough power to ring their customers' centage) of the current rating of the machine. telephones. This method also precludes the in- The sampling rate of the statistical ammeter stallation of unnecessarily large generators. is set by a 5 -cps square -wave generator. The Heretofore, there wasn't any simple method amplified output of the generator drives a refer- of measuring ringing current. The fluctuation ence clock counter at the rate of five counts per of the ringing current makes it difficult to meas- second. The same 5 -cps square wave is fed ure with conventional ammeters. An ammeter through electronic gates where it is amplified provides rather inexact information because its to drive two other counters. The level of the needle is hardly ever at rest. Operating company ringing current controls the operation of the engineers have used a special 20 -cps recording gates. The voltage drop across a resistor through ammeter in some central offices, but this involves which the ringing current passes is stepped -up a complicated analysis and interpretation. with a transformer, then rectified and filtered. This article concerns a new type of ammeter - The resulting dc signal turns the gates on or off. a statistical ammeter -which provides accurate There are three gates in the circuit. One is data immediately after a test run. It is portable adjusted so that whenever the ringing current and can be used with all types of ringing power equals or exceeds 75 per cent of the range -selector plants in the field. The meter samples the ring- setting, the 5 -cps square wave passes through the amplifier to a counter which operates at 5 counts per second. The other two gates are adjusted so that they will be turned on when the ringing current equals or exceeds 50 per cent and 100 per cent of the setting for the range -selector switch. The 75 per cent gate is connected per- manently to one amplifier and counter, and either the 50 per cent or 100 per cent gate may be connected to the other amplifier and counter. The range selector consists of resistors chosen so that the voltage drop at full load is about two volts on each range. The available ranges are 1/s, 1/4, 1/2> 1, and 2 amperes. These ranges were chosen to correspond to the full -load current ratings of 20-cps generators commonly used in ringing power plants. The different range resistors may be readily switched into the circuit with the range -selector switch. A variable resistor acting as an adjustable dummy load for the ringing plant and an ac milliammeter are used to calibrate the meter. The calibration procedure, although carried out on the 14- ampere range, is valid on all ranges. The meter operates on low power because all E. K. Ward of Pacific Telephone and Telegraph Co. of the active elements in the oscillators, ampli- adjusts a gate threshold on a statistical ammeter. fiers and gates are transistors. This also means

290 Bell Laboratories Record

www.americanradiohistory.com A short burst of electric current makes your telephone rind. Because' opera inq telephone companies must know how much current is required to ring hundreds of telephones like yours, they needed a combination of a pulse counter and an ammeter.

J. W. Osmun

A Statistical Ammeter

The usual method of measuring electrical cur- many cases where this information is either in- rent is with an ammeter. However, to measure accurate or unavailable. Further, there are so the amount of 20 -cps current required to ring many variations in service and customer tele- the hundreds of telephones served by a central phone habits from one central office to another office, we need a meter that "counts" rather than that such estimates are never wholly dependable. one that momentarily indicates the peak flow of The basic problem of determining the amount electric current. In other words, telephone com- of ringing current lies in the way the generators pany engineers must know how long as well as function. A ringing generator can deliver only how much ringing current passes through a cen- so much current before its output voltage drops tral office. below the satisfactory minimum. Thus, peak From traffic estimates, engineers can calcu- load current actually determines the size of a late with reasonable accuracy the expected ring- generator. The average ringing current, how- ing current for new central offices. With these ever, is much smaller than the peak current and estimates, operating companies initially can cannot readily be correlated with it. select generators with enough capacity to pro- Where ringing current drains are calculated vide ringing current for immediate needs as well according to probability theory, the generator as allowing for future growth. After a central output can exceed the generator rating by 20 office is in service for several years, however, per cent or more only one per cent of the time. it becomes difficult to determine how much elec- These momentary overload periods mean that a trical energy its generators provide. few telephones farthest from the central office Telephone engineers can estimate the cur- may not ring. This isn't serious because such rent required by determining the total number overloads last only a few seconds, and these of telephones served by a central office, the num- telephones will probably ring on the next ring- ber of these phones used in businesses and in ing cycle, usually six seconds later. Occasionally, homes, and the types of service. But there are central offices complain of alarms being set off be-

August 1961 2ß9

www.americanradiohistory.com three upward and three downward spikes cor- Nike -Zeus Fired responding to the three count signals that oc- curred, and we can count these as individual calls, From Underground Cell even though their original durations partly over- lap. Within some limitations, an adder with sev- A test firing of a Nike -Zeus missile from an eral branches followed by a differentiator gives underground cell at White Sands Missile Range the same results as several differentiators follow- was successfully completed last month. The mis- ed by a single adder, and it is much less expensive. sile's flight was guided by the Command Guidance A final trick from the bag has to do with dis- System developed at Bell Laboratories. This was play of the total count after it is obtained. the first firing from an underground cell of the Customarily, individual counts are displayed on advance configuration of the missile designed to four-digit electromechanical registers. These can counter ICBM's. be read manually or photographed. Traffic engi- Project engineers at the Laboratories said that neers want the electronic method of totalized the test was a significant achievement in the count to be displayed in this same familiar form. Nike -Zeus development program. It demonstrated Even a modern electromechanical register, how- that the Zeus system guidance equipment is fully ever, requires about 1 /20th of a second to count capable of controlling the missile. The Command a pulse, release, and be ready for a subsequent Guidance System used in the Nike -Zeus system pulse. Thus, if this counter is counting pulses at was developed at the Whippany Laboratory and a constant rate of 20 pulses per second, or 72 is manufactured by Western Electric at its North thousand an hour, it will be busy 100 per cent of Carolina works. the time, and unable to count additional pulses. In last month's flight, the missile performed satisfactorily over an extended range and all test Count Reducer objectives were met. The principal items tested To be accurate to within about 0.1 per cent on were in the Command Guidance equipment, de- randomly occurring pulses, such a counter must signed to track and guide the missile in complete- have its "busy" periods restricted to approxi- ly automatic operation. The underground launch mately 0.1 per cent of the time. It can, therefore, concept was previously demonstrated using a only be exposed to an average of 72 counts per modified version of the earliest Zeus missile de- hour -not enough for our purposes. Obviously, if sign. The underground cell is a prototype of the the desired electromechanical register is to be launcher which is planned for the missile defense useful, some sort of scaling -down or count- reduc- system. ing arrangement will have to be placed ahead of it. The Nike -Zeus system is being developed by the In serving telephone traffic over trunk groups, Laboratories under a Western Electric Company we know from probability theory that ten trunks prime contract with the Army Ordnance Missile can handle over 3000 times as many calls as a Command. single trunk for a "probability loss" of 0.001. Similarly, we can increase the capacity of a single electromechanical counter by preceding it with Western Electric Awarded a single electronic "scaler." This will permit every tenth call to operate the counter. With this UNICOM Contract modification, the capacity of the counter for 0.001 probability of loss is increased from an Western Electric Company has been awarded average of 72 calls per hour to an average of, a $19,101,500 contract by the U.S. Army for con- not 720 calls, but 213,000 calls an hour. Insofar tinued development of a Universal Integrated as lost calls are concerned this capacity is ample Communication System (UNICOM). for any proposed system application. The system will provide switching and terminal The foregoing set of simple principles is all facilities for a world -wide complex of voice, tele- that is required to permit pulse counting in elec- type, facsimile and other forms of communica- tromechanical systems. A more sophisticated tion. The network is expected to embrace a variety problem, of course, is the integration of these of transmission forms, including the use of artifi- basic techniques into a design where circuit cial satellites. complexity and cost are optimumly balanced Western Electric is prime contractor for against requirements of counting accuracy. Such UNICOM. The system is being developed by the a circuit is now in the final stages of design at Laboratories at its Whippany location. Subcon- Bell Laboratories. tractors will be used as appropriate.

288 Bell Laboratories Record

www.americanradiohistory.com mechanical dial system, when we try to produce shown, and the first time the input signal exceeds very short spikes by differentiating count signals a certain "trigger level" the flip -flop puts out a we run into trouble because these signals are cre- spike whose duration may be made practically as ated by closures of contacts on electromechanical long or short as desired. Moreover, the flip -flop relays and switches. Such contacts do not merely can be made to have a recovery or "dead" time close; they slam together, then quiver a while, following each spike which is also adjustable over making rather imperfect contact. They may even wide limits. Thus, the device gives a very brief bounce open again before settling down to pro- output spike when its input signal exceeds a set duce a nice steady "on" signal. Thus, a closure level and it ignores later input signal excursions may actually produce an `on" signal like that at caused by contact chatter or bounce. the left of part "b" of the diagram. If we use too Another technique, not necessarily restricted short an RC value, a register will get a spike for to pulse counting, is handy because it saves each short, preliminary closure or open and will money. This is the technique of "adding" the count each call several times. voltages representing two or more calls. As illustrated in part "e" of the diagram, four separate Single Downward Spike inputs connect through resistances to an output. The actual values of voltage or current in such a However, as indicated in part "c" of the figure, circuit depend on the connecting circuitry, but in we can select RC values to produce a long enough general the output signal at any instant has a spike so that the contact chatter is included as value roughly proportional to the sum of part of it. This leaves the second spike, the one the in- at put signals appearing the end of the "on" interval, as the only down- at that instant. Thus, if input signals such as those indicated ward spike accompanying the signal. We can come in on terminals 1 and 2, easily arrange to count only this one, and while the output signal is the sum of this is not evident in part "c," the downward these. This sum signal, sent through an RC "differentiator," spike may have a duration only a small fraction obtains the differ- of that of the original closure. ential of the sum signal. Note that we now have Actually, applying differentiating circuits to Techniques the input signals constitutes a form of frequency for counting electromechanical pulses. selection. That is, we have selected the medium frequencies representing the transition from "on" OFFI ON OFF to `off" (and vice versa), and have discarded the very low frequencies which give "body" to the input signal by representing its relatively long duration. In addition to rejecting the low frequency con- OFF1 ON tent of the input signal, it is also desirable to reject the very high frequencies which are picked up on the input leads. This is because such frequencies can easily radiate from one lead to another in central office cabling and we might find ourselves counting signals from circuits not attached to our input. For this reason, some simple arrangement to attenuate frequencies above 2000 cps is generally placed at a common point in -o the count -pulse path. FLIP Another technique for eliminating a tendency FLOP to score extra counts involves characteristics which may be given to a pulse -producing circuit called a "flip- flop." A flip -flop, as the name sug- gests, acts like a light switch -it is either "on" or "off." There are several varieties but the type which is useful in a pulse counter is illustrated in part "d" of the accompanying diagram. This flip - flop accepts an input signal such as the one

August 1961 287

www.americanradiohistory.com Arrangement of components for electronic pulse counter. Design bal- ances complexity and cost against required accuracy.

lasting for 100 milliseconds, onto one wire for miss only about one -half of one per cent of the counting, there are going to be many occasions counts, making our count 99.5 per cent accurate. when two or more count signals will partly coin- This is better than the accuracy ordinarily ob- cide and will appear merely as one long signal tained by present methods. rather than as two or more separate counts. Under Consideration of how short a count signal these conditions, we might expect that many should be, together with consideration of the cost count signals would be missed. In actual practice, of various ways of producing short counts, points from a third to a half of the counts would be up a second difficulty. About the cheapest and missed, depending on the specific counting mech- easiest way to turn a long signal into a short one anism used in the system. is to pass it through a so- called RC differentiator circuit, as indicated in the diagram appearing Shortening Signals on the next page. Obviously, each call signal has to be drastically At the left in part "a" of the diagram is an shortened if all are to parade single file into "on" signal representing a count to be shortened. one counter. But even with a very brief signal, After passing through the RC circuit, this signal occasionally two count signals will arrive so consists of a short "spike" corresponding to the nearly together that they cannot be distinguished start of the "on," and another "spike" of opposite as separate calls. Losses of counts from this cause polarity corresponding to the end of the "on." can, however, be kept insignificant if the count The duration of these spikes is determined by the signals can be shortened enough. values of resistance and capacity making up the For example, if we make each count signal only RC circuit. By selecting proper RC values, these We may 1 millisecond long and count an average of 20 spikes can be made very short indeed. thousand calls an hour, our counter is working select either the upward or the downward spike or "busy" for only 20 thousand milliseconds or for counting. We can feed many such short spikes 20 seconds out of each hour. Since there are 3600 per hour into a counter with few calls lost from seconds in each hour, the counter is occupied only over-lapping. 20/3600, or about one half of one per cent of the Unfortunately, however, we cannot use this hour. Thus, according to probability theory, we attractive scheme just by itself. In an electro-

286 Bell Laboratories Record

www.americanradiohistory.com Telephone traffic engineering and plant administration are based on knowing when and where telephone calls originale. To keep track of calls accurately, Bell Laboratories is designing an electronic pulse counter.

W. B. Callaway

The Art of Counting Calls

Customers of the Bell System originate about many scattered, individual registers. In this way, 300 million telephone calls a day. Keeping track the Bell System can reduce its 400,000 registers of this huge volume of business is a vast under- to a few thousand, and thereby save about a mil- taking, but it must be done if the Bell System is lion dollars a year. to handle properly its traffic problems. Only in The counting itself poses no particular design this way can the Operating Companies effectively problem. Electronic counters have long been avail- distribute their switching and transmission fa- able to handle many thousands of counts a sec- cilities to take care of the growth and other ond. And since the total number of calls to be changes in customers' telephone requirements. handled at any one place rarely exceeds about At present, the Operating Companies use ap- twenty thousand an hour, even a partly electronic, proximately 400,000 mechanical traffic registers partly mechanical counter can easily do the job. to count originating telephone calls. These regis- The real difficulty lies in trying to jam all the ters must be "looked at" about 100,000 times a individual count signals onto one wire without year. The time taken to read the numbers, when having them merge into an unrecognizable jum- translated into the wages of the clerks doing the ble. The situation is similar to one where every reading, amounts to nearly one -and -one -half mil- car in a city suddenly tries to go down a single lion dollars. Furthermore, such activity wastes street at the same time. human talents. The first part of this problem involves the In the interests of saving some of this clerical length or duration of each original count signal. expense, and, at the same time freeing personnel These signals were originally intended to operate for activities more in keeping with their abilities, electromechanical registers which require a Bell Laboratories has investigated the use of "long" signal to give the register time to func- electronic techniques for the counting function. tion. Therefore, most of these signals last more Electronically, it is possible to count signals on than 100 milliseconds. one central register which are now recorded on Now, if we merge 20,000 pulses an hour, each

August 1961 285

www.americanradiohistory.com pearance of a standard display on his "scope." If Command Guidance he does not get this indication within a few seconds, he can assume a malfunction exists some- Puts Tiros 111 in Orbit where in the system. In this case, he will normally request further information and instructions over his voice -radio telephone circuit. The pilot would The National Aeronautics and Space Adminis- make no attempt to service the equipment during tration's TIROS III weather satellite, launched a flight under any conditions. last month at Cape Canaveral, was directed into Prior to take -off, each aircraft will be given a its circular orbit by the Bell Telephone Labora- pre -flight check-out while still on the flight deck. tories Command Guidance System. A test set the size of a suitcase is plugged into This launch was the third successful use of the a jack on the side of the aircraft to verify that guidance system for TIROS (Television Infra - the communication system is in good working Red Observation Satellite) satellites and the fifth order. If trouble exists, the major unit involved time command guidance has been used success- can be identified in seconds. The defective unit fully in NASA's space research program. TIROS I can then be readily and speedily removed and re- was sent into an almost perfect a tested spare. At some convenient later circular orbit on April 1, 1960, placed by News of time, the defective unit is taken below decks to a and TIROS II, which is still trans- Space "test bench facility," where it is again checked in mitting meteorological data, was Research a simulated system. More detailed tests then are launched into a similar orbit on applied to indicate which of the various modules November 23 last year. TIROS II is responsible for the difficulty. When the module may be "turned off" now that its successor is in in trouble has been located, it is replaced by a orbit. tested spare and the major unit thus restored to The orbit of TIROS III is 509 miles at its apogee and 458 miles at its perigee. It passes 48 degrees north and south of the equator on each of its 100 -minute trips around the earth. The NASA- developed, Douglas -built THOR -DELTA vehicle used to launch TIROS II and TIROS HI was also used to place ECHO I, the passive communications satellite, and the Space Agency's Explorer X magnetic -field probe into their precise orbits. The other satellites and space probes in NASA's DELTA series will also be directed by the Bell Labo- ratories Command Guidance System. This guidance system, which Bell Laboratories developed for the Air Force Ballistic Systems Di- Digital -to- digital (right) and digital -to- analog vision for the TITAN I weapon system, is manu- converters of the airborne controlled terminal, factured by the Western Electric Company. Com- showing some of the typical plug -in modules used. mand guidance employs a special, ground -based digital computer- Athena -developed and manu- service. Normally no attempt will be made either factured by Remington Rand Univac. to locate or repair the defective component within The 285 -pound TIROS III satellite differs slightly the module at this time. Rather, the defective from its predecessors in that it has two wide - module will be returned to a similar land -based angle cameras instead of one wide -angle and one test -bench facility for ultimate test and repair. narrow -angle camera. This change was based on In addition to the tactical uses implied above, earlier experiments, which showed that more this data communication system is designed to valuable information for weather analysis was help direct aircraft in support of marine landing available from the wide -angle pictures. It also operations. It also will afford opportunities for a carries three infra -red sensors to measure sun - wide variety of applications in future military earth radiation relationships. and commercial communication systems. With TIROS III, like its predecessors, is designed to high- speed, digital data systems such as this, gather and record weather data which it spills out surface air communication methods are keeping as it passes over command and readout stations pace with the ever -increasing speed and complex- at Pacific Missile Range, California and Wallops ity of present -day Armed Forces aerial operations. Island, Virginia.

284 Bell Laboratories Record

www.americanradiohistory.com Right -hand section shows aircraft being directed interceptors are guided back to carrier and landed. toward their targets both from aircraft carrier Dashed colored lines represent coded radio mes- and from early warning type aircraft. At left, sages, while curved lines represent radar waves. modules. The digital -to- digital converter contains of the heat exchanger is illustrated in the dia- 37 similar, but not identical, modules. Printed cir- gram on page 282. cuitry is employed throughout the modules. Also, Although every effort has been made to insure wherever possible, each plug and jack element is reliability and complete freedom from trouble split so that, when connected, they make four during a mission lasting at least two hours, the parallel connections at each plug -jack junction. question still arises as to how so much compact This feature, in combination with self -aligning circuitry can be serviced in any reasonable time jack strips, helps insure good, permanent contact. when trouble does occur. The problem is espe- At the same time, it permits quick removal and cially difficult since such service will have to be replacement when required. handled by personnel who are not highly trained Portions of the airborne controlled terminal in the art. The answer to this question lies prin- will occasionally be subjected to temperatures cipally in the type of monitoring facilities and considerably higher than certain of its compo- testing techniques used. A brief description of nents can withstand as designed. Therefore, each how they work in the airborne controlled termi- major unit is self- contained in a pressurized, air - nal will serve as an illustration. cooled chamber. This chamber also protects the During a mission, the interceptor pilot is auto- components from moisture and air density varia- matically alerted by a visual or audible signal if tions which would otherwise affect the reliability for any reason he should fail to get a message of operation. The whole is cooled by a heat - every few seconds. In addition, if and when he exchanger, through which air circulates from the wants assurance that his equipment is function- aircraft's cooling system. The principle employed ing properly in all respects, he has merely to press for cooling by forced air circulation on both sides a button on the control unit and observe the ap-

August 1961 283

www.americanradiohistory.com system can provide two -way communication to as To avoid ambiguity in dispatching messages, many as 100 aircraft over a single radio -frequency each aircraft is given a discrete address. This channel. Each of these aircraft can be given a allows the message sent to a particular aircraft complete set of individual instructions once every to be received only by that aircraft. At the direc- few seconds. The two -way feature permits each tion of the pilot or the controller, the message addressed aircraft to reply with a comprehensive sent to the interceptor aircraft is either fed di- status -type message derived from information rectly into the autopilot (in which case only lim- generated aboard the aircraft. ited action is required on the part of the pilot) Two basic types of communication terminals are or "painted" on his radar scope in a pictorial dis- supplied : A control terminal located either on a play. In this case, the pilot "gets the picture" surface vessel or in an early warning type of instantly, without having to perform time -con- aircraft ; and a controlled terminal located in the suming translations or computing processes. In interceptor or fighter aircraft. Both types of ter- either case, at the completion of the control mes- minals are small and light- weight. They are also sage, a reply message indicating the aircraft's capable of encoding, decoding, and dispatching status is automatically encoded and dispatched the command and reply messages accurately at back to the control terminal. This status message the prodigious rate of some 10,000 words per is continuously generated from "own aircraft" minute. This accomplishment was realized by instruments and navigation devices, and thus re- making full use of all the latest pertinent tech- quires no expenditure of effort by the pilot. nological advances in the fields of data processing and digital data communication. More specifi- Service Conditions cally, the major units of the system, up to the A number of exacting requirements had to be very radio frequency signals, are completely met in developing this data link for the Navy, transistorized, and the system operates as a because of the service conditions under which it two -way device on a single line -of -sight radio would be used. The equipment not only had to be frequency channel in the UHF range. small and light- weight, but it had to function reliably and without failures at high altitudes, under extreme ranges of temperature, humidity, HEAT EXCHANGER FAN shock and vibration. Finally, it had to be so con- 1 1 structed that it could be operated and maintained by personnel who were relatively unskilled in INTAKE EXHAUST COOLING AIR servicing highly specialized electronic equipment. 1 1 ) For these reasons, great care has been exercised INTERNAL CYCLE EQUIPMENT PLENUM not only in the choice of components but in the choice of circuits and in the mechanical arrangement of the circuit components. CIRCULATING PATH EQUIPMENT TO For example: engineers carefully considered PRESSURIZED OF INTERNAL AIR BE COOLED CASE the relative merits of using transistors versus FAN tiny magnetic cores for the shift register ele- \1-'., \\\ ments in the data multiplex unit. They ultimately _' decided to use cores on the basis of the specific job to be done, taking into consideration the relative reliability, circuit complexity and the ease of identification and correction of a trouble con- dition if one should develop. In this instance, system reliability is concerned principally with the

' I intrinsic reliability of the component (transistor or core device) and the large number of components required. 111410, :..OVI.viqi / we have the specific i HEAT While said little regarding CHASSIS EXCHANGER design, circuit compactness in an ap- COOLING AIR equipment PLENUMPLE INTAKE EXHAUSTS plication such as this is a "must," and is perhaps BENEATH CHASSIS TO ATMOSPHERE well illustrated in the photograph appearing on Cofaicay showing heat -exchanger used to keep page 284. Both the digital -to- digital and analog - major communication units cool in operation. In- to- digital converters of the airborne controlled coming air comes from aircraft's cooling system. terminal are made up of individual functional

282 Bell Laboratories Record

www.americanradiohistory.com Carrier -based aircraft now Hit faster that the speed of sound. To direct the Nary pilots to their missions and get them safely back, hell Laboratories developed a nett' digital data communication system.

J. B. Bishop

A New Surface-to-Air Data Communication System

Speed in military aircraft has long been rec- developed a high- speed, surface/air digital data ognized as essential in maintaining air suprema- communication system for the U. S. Navy. The cy for our Fighter- Bomber forces. Development system, known as AN /USC -2, provides two -way work in this field made rapid strides during World communication under all weather conditions. It War II, and since that time jet aircraft have com- can be used to direct large numbers of Navy in- pletely superseded the propeller- driven type in terceptors simultaneously to their several targets this application. Air speeds equal to that of sound and, after they have completed their missions, put and above are now commonplace for these manned them in a homeward -bound traffic pattern which vehicles. will eventually bring them to a safe landing on To keep pace with these changes, our com- the deck of a carrier. munication facilities have also had to be radi- The facilities with which we are concerned in cally revised. Using conventional communication this article will be confined to the Data Link corn - systems, not more than a few of these high -speed munications portion of a completely integrated interceptors could be guided to their targets by a air defense system. Concurrent developments by director issuing voice commands over a single other organizations have furnished the high- radio channel. Communication requirements of speed computer, automatic radar- tracking de- aircraft in other phases of their mission, such as vices, displays, and other necessary supporting returning to base and landing, magnify this prob- facilities required at the control terminal. A great lem. Clearly, postwar surface /air communication deal of attention has also been paid to the matter systems required the same type of progressive of insuring a high degree of compatibility with development that enabled the aircraft industry to the surface /air communication facilities used by replace its slow-speed propeller type aircraft with other branches of our armed services, as well as the high -speed jets now in use. by those of other NATO countries. To meet this challenge, Bell Laboratories has This time -division digital data communication

August 1961 281

www.americanradiohistory.com Tite newest maser design to come from /fell Laboratories is one for a portable device that can be hand held and operated. Ils power supply/ is contained in a small case.

Portable Ruby Optical Maser Demonstrated by the Laboratories

be fired at ten A portable, pulsed -ruby optical maser, which 9 x 9 x 3 inches. The ruby can operates on less than one -tenth the input power second intervals and does not require cooling. of standard sized masers was demonstrated by Mr. Collins explained that the portable model purposes at R. J. Collins of Bell Laboratories at the Inter- had been developed for demonstration national Conference on Optical Instruments and lectures and seminars, but that it will probably Techniques in London last month. have other applications where space is limited. The light -weight optical maser can be hand- held and operated with an input power of only 128 joules from a portable power pack. Standard - sized ruby optical masers are several times larger and require hundreds of pounds auxiliary power equipment to News of of supply about 1500 joules of pump - Solid State ing power. Research The standard -sized maser uses a xenon flash lamp, wound spi- rally around the ruby rod, to stimulate maser action. The new maser, designed by J. W. Ammons of the Solid State Electronics Research Depart- ment, is made smaller and lighter through the use of a different type of flash tube enclosed in an elliptical reflecting cavity. A U- shaped, xenon flash tube is at one focal line of the elliptical cavity. The 2 -inch by 0.2- inch ruby rod is at the other focal line. Therefore, almost all the light from the lamp is focused onto the ruby rod. The ruby shows stimulated emission on only about 128 joules. The ends of the ruby rod are exposed. One end is partially reflecting and permits a very bright spot of light to emerge. The other end has two slits etched on it so that an interference pattern can be projected, demonstrating the coherence of the maser light. The power for the flash tube is supplied by a J. W. AIntnons sights along pot'rul m Maser hous- nine -pound power pack measuring approximately ing. The power supply hangs front his shoulder.

280 Bell Laboratories Record

www.americanradiohistory.com 36 350 noise temperature of O MEASURED SYST less than 20 degrees K. TEMPERATURE Such systems could be improved as much as 2 32 MEASURED SKY degrees K by reducing losses in the input wave - guide. Also, design changes in the maser feed 300 could reduce the noise by 5 or 6 degrees K. Thus, 28 it appears that a system noise temperature of z only 10 degrees K could be realized at the zenith. 250 > íWC 24 In this case, designers would be under pressure to V, keep the amount of physical structure to a minimum. For example, at such a noise temperature, 200 z the additional noise contributed by only 5 z feet of 6 kmc waveguide would halve the sensitivity, 41- 16 necessitating a very expensive doubling of the 150 required transmitted power.

12 Important to space communications is the vari- tY ation of the sky noise. The major variation comes ó 100 z z from changes in the content of the atmosphere; 8 rain, for example, produces changes. To study these variations, the Laboratories has initiated 50 4 a long -range investigation of sky noise and for this reason has replaced the temporary system with a permanent one. o 1 o 0 20 40 60 80 90 100 From the results of the sky noise study, we can ZENITH ANGLE IN DEGREES now evaluate the performance of a satellite com- System and sky noise vary according to zenith an- munications receiver. Low -noise equipment is gles. Theoretical calculations agree closely with now available, and at the proper frequency it will those obtained from actual sky measurements. be possible to send signals around the world with a minimum of interference. coupled from the noise lamp permits computation of the temperature of the over -all system. The sky noise measurements were performed Special Echo Issue of BSTJ on cool, clear winter nights so that the atmosphere Data gained in the balloon- bounce experi- would approximate the "standard" atmosphere. ment of Project Echo have now been sum - marized, and appear in the July issue of the A total of 250 individual measurements were Bell System Technical Journal. Titles include: made of the system temperature with the antenna Participation of Bell Telephone Laboratories directed at the zenith. Another set of measure- in Project Echo and Experimental Results, ments were then made with the antenna pointed W. C. Jakes, Jr. at different elevations. System Calculations, C. L. Ruthroff and W. C. Jakes, Jr. Since the sky temperature is proportional to 960 -mc, 10 -kw Transmitter, J. P. Schafer and the length of the antenna beam through the at- R. H. Brandt mosphere, it may be separated from the re- Receiving System, E. A. Ohm mainder of the system temperature figure A Horn- Reflector Antenna for Space Com- munication, A. B. Crawford, D. C. Hogg, through use of the observed variation with and L. E. Hunt zenith angle. The observed values agree well with The Dual Channel 2390 -mc Trarelieg -Wave the theoretical curves up to the point where the Maser, R. W. DeGrasse, J. J. Kostelnick, two -degree antenna beam intercepts the earth. and H. E. D. Scovil Standby Receiver System, L. U. Kibler The measured sky temperature agrees excel- FM Demodulators with .Negative Feedback, lently with the theoretical calculations. The mini- C. L. Ruthroff mum receiving- system noise that can be obtained Satellite -Tracking Radar, O. E. DeLange theoretically in an earth -based receiving system .961-me Lower-Sideband Up- Converter for pointing toward the zenith is the sky noise, and Satellite -Tracking Radar, M. Uenohara and H. Seidel is approximately 2.5 degrees K 6 kmc. at Prac- Antenna Steering System, R. Klahn, J. A. tically, of course, we must contend with other noise Norton, and J. A. Githens contributors. Nevertheless, the Laboratories ar- Boresight Cameras, K. L. Warthman rangement proved to have an over -all system

August 1961 479

www.americanradiohistory.com noise a few degrees Kelvin, any thermal radiation from K of noise, given the preamplifier a total the earth picked up by an imperfect antenna can temperature of 10.5 degrees K at the side of the represent a serious source of noise. For example, amplifier in the room -temperature environment. a parabolic antenna pointing at the zenith will The throat of the horn antenna, rectangular pick up thermal noise via its side and back lobes. in shape, tapers into circular waveguide which axis. Back and far side lobes emanating about 10 permits the antenna to rotate about the horn to per cent of the radiation of an isotropic or non - It then tapers back to rectangular waveguide directional antenna will pick up a noise contribu- enter the amplifier. The plumbing necessary to to tion of about 20 degrees K from the earth. permit transition of the signal from waveguide Thus, requirements for a low -noise antenna coaxial cable is attached directly to the input must include very low side and back lobes. Fortu- cable to reduce input losses. On the output side nately this requirement is compatible with other of the maser, the signal goes to an isolator, a desirable characteristics. Such an antenna will precision attenuator, another isolator, and then reduce the effects of man -made interference and to the crystal mixer, mentioned above, of a super- it will be efficient in the sense that its effective heterodyne receiver. area will approximate the actual area of the The technique for measuring sky -noise involved aperture. In addition, the antenna and its associ- comparison with a known noise signal from an ated feed structure must have low ohmic losses. argon noise lamp coupled to the input waveguide via a 26 -db directional coupling. When fired, the Horn -Feed Antenna noise lamp introduces additional noise into the to about 25 degrees K. The These requirements are satisfied to a high de- system equivalent change in the level of the output noise is meas- gree by a horn -feed parabola designed by Bell with a precision attenuator following the Laboratories for microwave relay applications. ured First, with the noise lamp off, the preci- First used in the Bell System's TD -2 radio relay maser. is set to some conveniently low system, it has played an important part in the suc- sion attenuator attenuation and the receiver gain is ad- cess of the Echo project. The radiation pattern of value of convenient reference level of detected this antenna shows its far side lobes at about 10 justed for a Then, as the noise lamp fires, the per cent of an isotropic antenna while the rear noise output. is adjusted to give the same detected lobes are at 0.1 per cent. Consequently, the total attenuation before. Knowing the excess noise earth noise picked up by the antenna with the noise output as beam pointed vertically is less than 2 degrees K. The amplifier used in the noise -measuring ex- 1000 800 is a -wave maser. This is a fi periments traveling 600 ink solid -state amplifier that uses a "slow -wave 400 M; structure" instead of a cavity for the signal. A comb -like structure acts to slow down the rf s z 200 wave. The active gain -producing material com- prising the maser is ruby. 100 Transmission lines leading from access termi- 80 W nals in the room -temperature environment to the O 60 cold maser are of material having low heat con- 40 ductivity. A "K band" waveguide connects the ce maser to a microwave "pump." At one end of the 20 a seal assembly is a plate which forms vacuum W for the helium dewar -a container for the liquid Ñ 10 amplifier at a low 8 helium needed to keep the z 6 > temperature. Ne In the middle of its tuning range (5.65 kmc), N 4 the traveling -wave maser has a gain of 35 db at a bandwidth of 25 mc. This gives the amplifier 2 sufficient gain to override the noise of a crystal 20 40 mixer following it in the circuit. The maser it- 0.1 0.2 0.4 0.6 LO 2 4 6 8 10 self has a noise temperature of 2 degrees K at its FREQUENCY IN KMC PER SECOND frequency; low-temperature input. The coaxial connection Galaxy noise decreases with increasing sources increases. on the input side, however, contributes 8.5 degrees noise emanating from earth

Bell Laboratories Record 278

www.americanradiohistory.com DeGrasse, E. A. Ohm, and the authors. With a munications, and since the sun and moon occupy high -efficiency horn -reflector antenna and low - only a small angle, they are easily avoided. How- noise maser, measurements were obtained from ever, the galaxy is a widespread source with vary- which it is possible to predict accurately the lowing intensity. Fortunately, this noise from the est noise "temperature" that can be achieved with galaxy decreases with higher frequencies, and its present systems. It is interesting to note that this interference contribution becomes negligible antenna -maser combination is but a scaled -down above 5 kmc. version of the one later used for the Echo satellite Terrestrial noise comes from the oxygen and program. water vapor in the earth's atmosphere, which ab- In discussing the sensitivity of a receiving sorb radiation at microwave frequencies and system, it is convenient to refer all sources of consequently act as sources of thermal noise. This electrical noise to one place -where the signal absorption, and hence the thermal radiation, also enters the system. In this way the real system depends on frequency, but decreases with decreas- can be replaced by an artificial noiseless one with ing frequency, becoming on a dry day, very low a noise generator at its input whose magnitude below 5 kmc. Moreover, the effective noise tem- of output noise represents the total noise in the perature of the atmosphere depends on how much actual system. In general, the noise power can be atmosphere the antenna beam goes through. expressed by the product of the absolute tempera- Hence the noise temperature will vary with the ture of this fictitious noise source and the band- elevation angle of the antenna, being a minimum width of the system in cycles per second. It is when the antenna is pointed at the zenith. This then possible to refer, instead of to noise "power," angular dependence is shown in the figure on the to noise "temperature" which has more direct next page. significance in a very low -noise system. Further- Theoretically, then, the total sky noise (galaxy more, just as noise powers can be added, so can plus atmosphere) exhibits a broad minimum at noise temperatures. Consequently, the effective about 5 kmc. This is the primary reason for per- noise temperature of a system consists of the forming the experiments near this frequency. sum of the noise temperatures of its constituent Communications system connecting one point to components, all referred to the input. another on earth, where signals travel parallel Any -bound earth receiving system will pick to the earth, will pick up about 100 degrees K up noise from the thermal radiation of the sky noise from the atmosphere and the ground. Thus, and this will set a lower limit to the sensitivity of receiving systems with noise temperatures of 20 such a system. This noise originates both ex- degrees K or less are likely to find their main traterrestrially and terrestrially. applications in satellite or space communications Many of the extraterrestrial sources of noise at frequencies of around 5 kmc, and where the interest the radio astronomer. In most cases, antenna elevation is 10 degrees or more above however, only the sun, moon and our own galaxy the horizon. constitute sources of serious interference to corn- Since electrical noise from the sky may be only

SIDE OR DIRECT SKY NOISE TS =2.5'K NNA WAVEGUIDE - AND TA= 1': COUPLER LOSS TWO = 2K

NOISE TOUT = 58,500 LAMP MASER OUTPUT "26 DI NOISE COUPLER

TWM NOISE Tm = 10.5'K Gam = 35 DB

How horn -reflector antenna connects to traveling- equipment by low heat- cowl nrti i'it y lines. K -band wave maser. Cold maser is connected to rest of waveguide connects the maser to microwave pump.

August 1961 277

www.americanradiohistory.com Research in space communications lias uncovered some problems heretofore of minor interest to communications engineers. One of these is noise from the atmosphere. Studying it has required extensice research into the use of antennas and amplifiers.

D. C. Hogg and H. E. D. Scovil

Measuring the Sky's Electrical Noise

the most One of the great obstacles to be hurdled in de- Heretofore, the amplifier was by far signing an efficient communications system is to serious contributor of noise in a communications But overcome the effects of electrical noise. Electronic system operating at microwave frequencies. direct devices, even when at rest, abound in noise the ultra low-noise devices now available "generators" in the form of random fluctuations us to examine the entire system for its separate of the atoms in their materials. And when these noise contributors. Masers now being developed, devices operate, the agitated atoms generate a for example, are so sensitive that even the trans- might great deal of noise while amplifying the signal. mission line connecting one to an antenna to the Today, we want to send signals over longer and add several times the noise intrinsic longer distances, to communicate via artificial maser. Furthermore, the antenna, the antenna satellites. The successful balloon- bounce experi- feedhorn, and even the sky itself become im- ment of Project Echo (RECORD, September, 1960) portant contributors. first is but the first step in this method of transmis- Before any disease can be cured, it must of sion. Thus, it seems imperative to find ways to be diagnosed. Similarly, to learn the abilities generate and amplify microwave signals with as a long -range communications system, such as little noise as possible. Fortunately for this re- provided by artificial satellites, we first need to quirement, new devices have recently arrived on learn about the inherent noise in the system. For the communications technology scene, such as the this reason, a series of noise measurements was Holm- parametric amplifier (RECORD, October, 1959) and conducted on top of Crawford Hill, near the by R. W. the maser (RECORD, July, 1958; May, 1960). del, N. J. location of the Laboratories,

Record 276 Bell Laboratories

www.americanradiohistory.com with its wrapped tape is easily produced in long lengths, which are simply cut to size for assembly into an array. Also, in reading out a bit, the reversal of the magnetic flux has a transformer action, since the magnetic flux wraps around the center copper conductor several times. A small interrogating voltage produces a larger output voltage. And, like core memories, the twistor is a high -speed device which can store large amounts of information compactly. In such a twistor memory four bits are usually stored per inch of wire. In a variation of this device, a twistor becomes the sensing circuit. The actual information is stored in an array of small bar magnets mounted in a card. That is, at each point in the X -Y coordinates, there is a choice of placing a magnet on the card or of omitting it. If a magnet is placed at a coordinate position, its external field inhibits the switching of the twistor segment. Thus, applying currents to the twistor in effect Lee H. Gallaher, Electronic Central Office Devel- interrogates the data stored on the card. When opment Department, adjusts lenses on servo sys- the segment cannot be switched, this indicates tem of experimental flying spot store memory. a magnet on the card; switching indicates no magnet. The structure is a non -destructive those for which magnetic materials are used. memory, and might be called a "card -changeable" Another memory technique now under investi- device. Small cards of magnets may be removed gation is the phenomenon of superconductivity. and inserted in somewhat the same manner as Certain metals, when refrigerated to tempera- that used with photographic plates in the flying - tures near absolute zero, lose all their electrical spot store. The use of many magnet cards may resistance. Further, with the application of a make the capacity of the memory comparable to magnetic field, they can be rapidly switched be- that of the flying -spot store. tween this superconducting state and the state of To this point, electronic memories have been normal conductivity. Devices based on this phe- described in terms of magnetism, electrostatic nomenon, called "cryotrons," can store very large storage, and photographic exposures. But these, amounts of information in a small space, and their of course, by no means exhaust the physical switching speed is very fast. Refrigerating to phenomena that could be used to write, interro- extremely low temperatures is a problem for some gate and read out communications information. applications, but this should become less serious Three additional possibilities will be mentioned as low- temperature technology advances. briefly here. The first is a variation on the idea All the devices discussed in this article have of the flying -spot store. Instead of the screen of been referred to several basic requirements of a cathode -ray tube, we might use an electro- electronic switching. Work has been directed luminescent screen with a matrix of wires on both toward temporary and permanent memories, and sides. Applying signals to two of the wires toward reading the information either serially or would then generate an electroluminescent spot in parallel. The concept of random access to the at an X -Y position defined by the intersection of stored data is also important. Further, the em- the two wires. phasis is on very small pieces or segments of Ferroelectric materials have also been investi- memory materials and on very high switching gated for use in memory systems. When a voltage speeds. Ultimately, such requirements are aimed is applied to a ferroelectric crystal, the surface of at the design of the communications systems of the crystal accumulates an electrostatic charge. the future -equipment that will occupy only a Since the phenomenon here is analogous to a small amount of floor space but will perform change in magnetic flux, the proposed methods existing services better, and with a greatly in- of using ferroelectrics are quite similar to creased capacity at the lowest possible cost.

August 1961 275

www.americanradiohistory.com memory is de- Y row. These two half -currents are insufficient memory is desired. In some, the the output cir- to switch the cores in either the column or the signed to be regenerative. Since memory, it is a fairly row, except for the particular core at the one cuits know what was in the to put it back in again X -Y point which receives both currents. This straightforward procedure then reverses its direction of magnetization. after a readout. core each In reading this bit out of the core, coincident cur- In another variation, the "transfluxor," central hole are similarly used. If the two half -currents core not only has the main or rents smaller hole on a core that is already in the for storage but also has another are impressed through these magnetic state that would be produced by these through its wall. Wires passing presence of stored currents, little change results. But if the core smaller holes can detect the changing the magnetic state is in the opposite state, the core is switched, and information without memory constructed with consequent reversal of flux induces a voltage of the large hole. A the readout memory. pulse in a "sensing" wire. Read -out thus con- such cores is a non -destructive the "inhibited fluxor," sists of sensing the pattern of voltages and no- In still another variation, have three holes each. voltages from the memory. small pieces of ferrite are used for read- For this type of storage, it is evident that the Wires through two of the holes flux change around the two directions of magnetization must be un- ing and writing while the may be written ambiguous. In magnetic terms, this means that third hole is sensed. Information of currents in the hysteresis loop of the magnetic material into the core by the coincidence fluxor can be switch- should be rectangular with a well defined "knee." the two holes. The inhibited speeds using fairly simple Magnetic fields below the knee or threshold should ed at extremely high of storage. produce no switching, while fields slightly larg- schemes for selecting the points the knee should completely reverse the er than The Twistor Memory magnetic flux. After the currents are removed, for new and improved memories has the flux density should stabilize at a remanent The search the "twistor " -an- induction not too much less than saturation. led to a new device known as magnetic -core memory, but Various ferrite materials (RECORD, April, 1957) other relative of the Its basic principle of op- have been particularly useful for such applications. one that has no cores. a magnetic field produced It is also evident that with this type of core eration is that circular and a linear magnetic arrangement, the act of reading the information by a current in the wire combine to produce a removes it from the memory. Read -out is "de- field applied externally structive," and the memory is termed "erasible." helical magnetization. a small magnetic tape For this reason, a number of variations have been In one form of twistor, a copper conductor. tried for applications where a non- destructive is wrapped helically around Unlike the core memory, in which the magnetic flux is wrapped around itself in closed ring, the twistor has its magnetization along a length of the wrapped material -over several turns of the tape. Thus, successive bits are stored along a length of twistor wire. The circular magnetic field is produced by a current through the copper conductor, and the linear field is produced axially by a coil surrounding the particular segment to be magnetized. The method of inserting a bit bears some similarity to the method used in the core memory : a circular field from the current in the copper conductor, by itself, is not sufficient to switch the magnetic flux in a segment of tape. Neither is the current in the coil producing the axial field. Where the two coincide, however, the segment of magnetic tape is magnetized by the resultant field. Section of tici,titor magnet )ienroril .rritlt three The twistor structure has a number of ad- cores and word coils, intersecting eight twistors. vantages. In particular, the copper conductor

274 Bell Laboratories Record

www.americanradiohistory.com relatively permanent, and persists as long as the by their ability to store large amounts of infor- film remains in the memory system. mation in a small volume, by their simplicity of This system is called a flying -spot store because equipment, and by their relatively low cost per a moving spot of light appears on the face of a bit of stored information. Their chief handicap cathode -ray tube. With an arrangement of lenses, for some applications is that the memory surface light from this spot is simultaneously directed at or the reading device must move to recover the a number of photographic plates. If it strikes a memorized data. That is, for instance, the reel plate at a transparent area, it passes on through of tape must unwind or the drum must rotate in and, with another lens, is focused onto a photo - order to select a particular section of data. Read- multipler tube. An output from the photomulti- out therefore may take a few thousandths of a plier tube and its associated amplifier indicates second or longer. a transparent area on the film, and no output, of Such "slow" operation is not inherent in mag- course, indicates an opaque area. Thus, the flying netic memories, however. One method that has spot on the face of the cathode -ray tube in- proved extremely successful for high -speed mem- terrogates the photographic plates, and words ories is to fabricate a magnetic material in the are read out from the photomultipler circuits. form of "doughnuts" or cores. Several thousands of these are usually assembled in a coordinate Large Memory Capacity frame of wires. The wires are threaded in the X Flying-spot stores now in use have four large and Y directions through the cores at each inter- photographic plates, each having 17 sections for section. With a magnetic -core memory, words storage of binary bits. One section has nearly may be selected at random, and the read -write 33,000 spots, and the total memory capacity is cycle may be only a few millionths of a second. therefore 4 x 17 x 33,000, or over two million bits. Cost per bit of storage capacity is often quite The entire group of films can be interrogated very small. rapidly as the electron beam moves over the face Information is stored according to the direc- of the cathode -ray tube. Any word of information of magnetization in a core -that is, the tion can be read out in about 2.5 millionths of a magnetic flux may be directed clockwise or second. counter -clockwise around the ring. Once "switch- As does the barrier -grid tube, the flying -spot ed" in either direction, the core remains mag- store requires very accurate positioning of the netized after the magnetizing force is removed. beam. For this purpose, extra optical channels To write information into a core, a certain and corresponding photographic plates are in- value of current is necessary. Half this current corporated into a closed -loop servo system. Any is applied to a wire passing through an X column error in the position of the beam is fed back to of cores, and the other half to a wire through a make the required correction. The development of this servo system for the flying -spot store is a major contribution to the art of designing high - capacity, high -speed memories. If we now turn to others of the basic physical and electrical phenomena that can be used for storing information, certainly one of the most important and most useful is magnetism. And again, the fundamental idea behind most magnetic memory systems is the presence or absence of the phenomenon in a small section or area of the storage material. On a magnetic tape, for example, a bit of information is specified by whether or not a small area of the tape is magnetized in a particular direction. Similarly, the magnetic material may be plated on a cylinder to form a magnetic drum. Individual bits may be entered by magnetizing areas in rings on the surface of the drum. Many bits of information are stored in the bi- Magnetic tapes and drums are characterized state magnetic cores strung on this matrix of wires.

August 1961 273

www.americanradiohistory.com .1 new concept in automatic manufacture Of electronic components combines statistical quality control with mechanized processing at the Western Electric Company.

Resistors for Nike -Zeus Made On a Completely Automated Line

Carbon deposited resistors- untouched by hu- At the coating machine, a control device regu- man hands -are being turned out at the rate of lates the speed of the core through the three one every three seconds by a unique computer - separate chambers of a horizontal furnace. A controlled production line. The 110 -foot long fa- horizontal three -roller support system rotates it cility, recently shown for the first time, was de- en route to assure a uniform coating. In the heat- veloped by Western Electric Company's North ing areas of each chamber, a temperature of over Carolina works to meet stringent military re- 2000 degrees F. decomposes methane gas to form quirements for the Nike -Zeus anti -ICBM. crystallizing carbon on the core. After cooling, The heart of the resistor is a tiny ceramic the core is sent to an inspection station. core produced from processes developed by Bell The first inspection station forms a feedback Laboratories. Fed automatically loop from the coating furnace through the com- News of from a hopper at the beginning puter. Here the coated core passes between four of the line, the core moves The Bell along probes using the Kelvin bridge principle. The a conveyor belt to the coating results of this inspection, after being measured System machine at a speed controlled to against the programmed requirements, form the within an accuracy of 0.1 per cent. basis for feedback control of the furnace. The line, which Western Electric explained is The core next goes to a terminating machine the first completely automated process to manu- which sputters a gold contact over each end. First, facture any discrete electronic component, con- it is fitted with a mask which holds and protects sists of eleven stations, all tied into control by a its center. (The computer programs proper mask general purpose computer. Feedback of process sizes for four different sizes of resistors.) The data from three key points along the line permits mask is loaded onto a vacuum station and cov- rapid closed -loop operation. ered with a bell jar. After the jar is pumped to The process begins with the deposit of carbon a vacuum and backfilled with argon gas, the ends on the ceramic core. The core then goes succes- of the core protruding from the mask are sput- sively through inspection, termination, capping, tered with particles from a gold cathode. The spiralling to value, second inspection, molding of sputtering lasts for almost a minute, depositing a protective case, marking, leak inspection, final a layer of gold approximately ten millionths of inspection and packing. an inch thick.

August 1961 295

www.americanradiohistory.com Western Electric engineers inspect computer- production line. Cathodes in bottom of each bell controlled terminating machine in the automatic jar sputter layers of gold on ends of resistor cores.

296 Bell Laboratories Remrd

www.americanradiohistory.com At this point, the ceramic core is a resistor. stamps the wattage, resistance value, production The capping station now attaches wire leads to lot number and date on the encapsulated resistor. each end. The wire leads are first attached to Servo drives controlled by the computer can set tiny hexagonal caps of gold -plated brass by an any one of 45 different code numbers with over automatic percussion welding machine outside the a million and a half permutations of code and line. The capping machine uses two different resistance value combinations. capping assembly heads and three different sizes The final inspection station, a feedback control of caps for the four resistor wattage sizes. The point, resets the preceding inspection station to computer controls both the switching of the as- compensate for shifts in resistance value caused sembly heads and the changing of the cap sizes. by the heat of encapsulation. The helixing machine, the next station on the The acceptable resistors are then packed in line, gives the resistor a precise value. To do this, lots by the packing machine and inserted into it cuts a spiralled groove along the carbon film of Styrofoam blocks -the final process in this fully the core by rotating the properly-chucked resistor automatic operation. against a diamond- impregnated wheel. A com- The heart of the control equipment, the digital puter- controlled bridge monitors the cutting. The computer, has an over -4000 -word magnetic -drum bridge's control servos balance when the desired memory. Western Electric engineers extensively resistance is reached. During cutting, the bridge redesigned its input and output circuits to con- also senses any chipped or unevenly coated cores trol the programming, setup and feedback control and rejects them. It also rejects any resistor that of the individual machines. To produce resistors reaches value before 75 per cent of its length is with four different wattages and a huge number used or fails to reach value in its full length. The of resistance values the computer performs four helixing is done "dry" and continues until final basic functions: resistor value is reached; this eliminates the hand - rubbing final adjustment method previously used. It programs production requirements for a The feedback control and self- correction of the month, completely scheduling and arranging the work sizes and helix lathe bridge is based on a statistical quality according to the power resistor control analysis of values inspected at the second values to be produced. inspection station. A wheatstone bridge, set to It analyzes control statistically from data the desired nominal resistance value of each re- plotted at three critical points in the automated sistor lot by the computer, precisely measures the process and determines if a trend is developing. resistance. Off -balance voltage is digitized and It formulates information to detect any drift fed back to the computer, which then calculates from manufacturing tolerances and, if such a the desired correction and adjusts the helixing drift develops, calculates new setup information machine setting. for the appropriate station from previously stored The resistor is then fitted with a precured data. epoxy shell and two partially cured epoxy pellets It provides the initial setup of wattage size and placed between two resilent rubber chucks at eight machines and resistance value at six. before it enters the encapsulating machine. The resilient chucks prevent trapped air from causing In missile systems, deposited carbon resistors leaks in the finished resistor. are required by the millions and they must work After they are cured for about 15 minutes in under extreme conditions of cold, heat, humidity, a 300 degree F oven, the partially cured pellets vibration and shock. In certain defense equip- soften and form an effective seal with the pre - ment, a failure rate of no more than one per 200 cured shell. To terminate the curing process, the million hours -about 23,000 years -of operation resistors are passed through cooling water jets. is permissible. Detection circuits within the machine reject any This reliability requirement is mainly respon- resistors with missing shells or missing pellets. sible for the development of the automatic line. Moving along the conveyor belt from the oven, Until now, individual precision products have the resistor is immersed in a hot water bath been made by manual or semi -automatic processes which contains a wetting agent. This prevents and they were subject to contamination from surface bubbles from clinging to the resistor handling and other shortcomings of human con- body. The heat expands the air inside the capsule, trol. With advanced automation, there is no and a series of ten photoelectric cells watch for contamination problem, production is vastly in- air bubbles which would indicate a leak. creased and a level of reliability is attained that A computer controlled marking machine now manual methods cannot match.

August 1961 297

www.americanradiohistory.com Trial Begins On PCM Transmission System

A major advance in telephone transmission The Ti, forerunner of PCM systems that may called pulse code modulation, or simply PCM, was eventually stretch across the country, was de- put on trial early last month over Bell System signed to serve metropolitan area routes of up to customer's telephone lines between Newark and 25 miles. Other systems that increase the message Passaic, New Jersey. The experimental PCM sys- capacity of wires and cable have been used for tem, the T1 system, was developed by the Labora- many years by telephone companies, but they tories at Murray Hill and Merrimack Valley. have been economically feasible only over fairly The Ti system does not send a continuous long routes. voice signal as do conventional systems. Instead, The new system is expected to be particularly it takes samples of the speaker's voice very rapid- useful in large cities like New York, where con- ly -about 8,000 samples a second -encodes them gestion below ground has often made it difficult and sends out code pulses. At intervals of approxi- to find room for additional telephone conduits. To mately 6,000 feet along the cable route, the coded increase telephone facilities along a route, tele- signals are regenerated and sent on to the next phone companies will not have to dig up city repeater point. At the other end the pulses are streets. Instead, they may increase the capacity of decoded and the original voice signal is recon- their existing cables by installing T1 terminals in structed. telephone buildings at each end of the route, and The system sends pulses over a cable pair at the Ti repeater equipment in manholes or on poles rate of one and one -half million pulses per sec- along the way. ond. This rate permits the codes of many differ- Starting early next year, Western Electric ent voices to be interlaced on the same wires, thus Company will manufacture Ti for general use increasing their message -carrying capacity. throughout the Bell System.

H. E. Brown, aborc, and M. S. C 11ì1ì off check out Ti System equipment in a Newark, N. J. manhole.

298 Bell Laboratories Record

www.americanradiohistory.com Announcement System Makes New York Debut news in brief The 9A Announcement System, developed at Bell Laboratories (RECORD, February, 1959) was recently put in service for the A. L. Stott Elected Mr. Stott joined the Bell System New York City Report. The re- in 1929 as a clerk in A.T. &T.'s port consists of a one- minute re- A.T.&T. Vice -President chief statistician's division. Later corded message on matters of toll Alexander L. Stott, A.T. &T. he became supervisor of general municipal interest. In comptroller was recently elected studies in the Long Lines Depart- emergencies, such as transit tie - as- fires a vice -president of the company. ment and following that, staff ups, and severe weather con- sistant in the Treasury Depart- ditions, bulletins will keep New ment. He was appointed assistant Yorkers posted on up-to- the -min- comptroller in 1949 and assistant ute conditions. The message is treasurer in 1952. He was elected recorded at a master control cen- treasurer of A.T. &T. in December ter and fed first to a Manhattan 1952 and comptroller in the fol- central office and then through lowing year. a network of direct telephone Mr. Stott received his A.B. de- trunk lines to central office sub - gree from Harvard University in centers in other boroughs. 1929. He served with the U.S. The service, similar to the tele- Navy from 1942 to 1945 and at- phone company's time and weath- tained the rank of commander. er announcements served by the He is a trustee of the Committee 3A Announcement System (REC- for Economic Development, a ORD, November, 1939), was bought member of the Controllers Insti- by the city because police, trans- tute of America and a vice presi- it and other city information dent of the New York City Con- channels were overloaded during A. L. Stott trol. Hurricane Donna last fall.

waveguide science. It was at New Building for Crawford Hill this site that Karl Jansky did his memorable work on the The Laboratories has an- modate about 120 people who world's first "radio telescope." nounced plans for the construc- are now working in the old The scientists and engineers tion of a one -story laboratory frame structure. who will move to the new build- building on its Crawford Hill The old buildings were a cen- ing will continue their radio property in Holmdel. Construc- ter of research in radio and and waveguide research there. tion will begin this summer and plans are for the building to be completed in early 1962, at the same time the new development center off Crawford- Everett Road is opened. The new building will replace the existing frame structures in front of the development center. These frame buildings are a research outpost which began in 1929; they will be razed to make room for site grading and roads at the new center. The new building will be of masonry, and will have two one - story sections connected by a wing. It is planned to accom- Artist's sketch of the proposed new building for Crawford Hill

August 1961 299

www.americanradiohistory.com TALKS

Following is a list of speakers, titles and places of presentation Considerations of Noise Per - for recent talks presented by members of Bell Laboratories. formance, I.R.E. PGMT &T Na- tional Symposium, Wash., D. C. Felder, H. H., and Osgood, D. T., SEMICONDUCTING COM- Bennett, W. R., Jr., Radiative A Large Scale Four -Wire POUNDS CONFERENCE, Lifetimes and Collision Trans- Switched Communications Net- Schenectady, New York. fer Cross -Sections of Excited work for Military Communica- Atomic States, Quantum Elec- tions -The Transmission Plan, Dietz, R. E., and Thomas, D. G., tronics Conf., Berkeley, Calif. Excitons and Absorption Edge Fifth National Symposium on Bennett, W. R., Sr., Recent Trends Chi- of ZnO. Global Communications, in Communication, Weekly Col- cago, Ill. Hutson, A. R., Piezoelectricity loquium Elec. Engg. Depart., and Semiconductivity in Wurt- Flanagan, J. L., and Bird, C. M., University of Minnesota, Min- Minimum Phase Responses for zite and Zinc Blende Semicon- neapolis, Minn. ductors. the Basilar Membrane, Sixty - Bhuta, P. G., Nonsteady Two -Di- Morin, F. J., Halides, Oxides, and First Meeting of Acous. Soc. mensional Jet Mixing Theory Am., Phila., Pa. Sulfides of the Transition Met- f or Laminar and Turbulent L., David, E. E., Jr., als. Flows, Stevens Institute of Flanagan, J. B. J., Binaural Thomas, D. G., Excitons and Band Technology, Hoboken, N. J. and Watson, in CdTe. Lateralization of Cophasic and Splitting Bird, C. M., see Flanagan, J. L. Thomas, D. G., see Dietz, R. E. Antiphasic Clicks, Sixty -First Brattain, W. H., Brattain on Meeting of Acous. Soc. Am., Semiconductors, The Norfolk Phila., Pa. College of William and Mary, SYMPOSIUM MOLECULAR of ON Norfolk, Va.; Sigma Xi Lec- Fulda, S. M., Some Aspects STRUCTURE AND SPECTROS- in. Indus- ture, Stevens Institute of Tech- Systems Engineering COPY, Ohio State University, Res. nology, Hoboken, N. J. try, Air Force Cambridge Columbus, Ohio. Labs., Bedford, Mass. Brattain, W. H., Seminar on the Adamson, A. W., see Liehr, A. D. Surface Properties of Semicon- Georgopulos, S. G., 5 Type Artifi- Liehr, A. D., The Three Electron ductors, Chem. & Phys. Faculty, cial Larynx, Irvington Sertoma (or Hole) Cubic Ligand Field The Norfolk College of William Club, Indianapolis, Ind. Spectrum. and Mary, Norfolk, Va. Germer, L. H., Low Energy Elec- Liehr, A. D., Perumareddi, J. R., Collins, R. J., Optical Maser, Op- tron Diffraction Studies of Ad- and Adamson, A. W., Spectra tical Soc. Am., Chicago, Ill. sorbed Gases, Phys. Colloquia, of the One and Two Electron Connolly, R. A., see DeCoste, J. B. Rutgers University, New (or Hole) Transition Metal Cy- Courtney- Pratt, J. S., A Note on Brunswick, N. J.; Res. Depart., anide Complexes. the Possibility of Photograph- United States Steel Corp., Mon- Perumareddi, J. R., see Liehr, ing a Satellite Near the Moon, roeville, Pa. A. D. The Ultimate Sensitivity in Gianola, U. F., Looney, D. H., Porto, S. P., and Wood, D. L., Photography Conf., London, Mum, A. J., and Ruff, J. A., The Optical Maser as a Raman England. Large Capacity Card- Change- Source. David, E. E., Jr., see Flanagan, able Permanent Magnet Twis- Snyder, L. C., The Interaction of J. L. tor Memory, Symposium on Pi Electrons with C -C Bond DeBenedictis, T., see Hansen, R. H. Large Capacity Memory Tech- Length Displacements in Pseu- DeCoste, J. B., and Connolly, R. niques, Wash., D. C. doaromatic Molecules. A., Plastics Versus the Ele- Gibbons, D. F., Magneto Acoustic Wood, D. L., Electronic Spectra ments, Soc. of Plastics Ind. Measurements of the Fermi of Divalent Rare Earth Ions in Meeting, N. Y. C. Surface of Zinc and Cadmium, CaF,. Deutsch, M., Bargaining `Face': Princeton University, Solid Wood, D. L., see Porto, S. P. Some Experiments in Bargain- State Colloquium, Princeton, ing Behavior, Nineteenth Na- N. J. tional Meeting of Operations Gibson, W. M., Structure in the OTHER TALKS Res. Soc. Am., Chicago, Ill. Kinetic Energy Spectrum of Anscombe, F. J., Testing to Estab- Elkind, M. J., Gold -Plating Elec- Fragments from the Thermal- U'ss lish a High Degree of Safety tron Devices, Western Electric Neutron- Induced Fission or Reliability, Inst. Math. Sta- Interworks Finishers Conf., Co- Oak Ridge National Lab., Oak tistics Eastern Regional Meet- lumbus, Ohio. Ridge, Tenn., Los Alamos Sci- ing, Ithaca, N. Y. Engelbrecht, R. S., Elementary entific Lab., Los Alamos, N. M.

300 Bell Laboratories Record

www.americanradiohistory.com Gibson, W. M., and Miller, G. L., ance front Matrix Electrolznni- Meeting of Acous. Soc. Am., Charge Collection in Semicon- nescence Screens in Optical Phila., Pa. ductor Particle Detectors, Nu- Readout Applications, Large Schwartz, N., Tantalum Thin clear Electronics Conf., Bel- Capacity Memory Techniques Film Components and Inte- grade, Yugoslavia. for Computing Systems Sym- grated Circuitry, I.R.E. Prof. Glaser, J. L., Radio Communica- posium, Wash., D. C. Gp. on Component Parts, tions via Satellites, Wichita, Looney, D. H., see Gianola, U. F. Wash., D. C. Kan.; St. Louis, Mo.; Cincin- Matthias, B. T., Ferromagnetism Smith, W. L., see Spencer, W. J. nati, Ohio. and Superconductivity, IBM Spencer, W. J., and Smith, W. L., Glaser, J. L., The Planning of Conf. on Superconductivity, Precision Crystal Frequency Commercial Satellite Communi- Yorktown Heights, N. Y. Standards, Frequency Control cation Systems, Chicago, Ill. McLean, D. A., American Devel- Symposium, Atlantic City, N. J. Gnanadesikan, R., Graphical Anal - opment in Miniature Capaci- Tebo, J. D., The Nike Missile ysis of Multi- Response Experi- tors, I.E.E. Symposium, Lon- Family, Old Guard, Summit, mental Data Using Ordered don, England. N. J. Distances, I.M.S. Meetings, Miller, G. L., see Gibson, W. M. Tebo, J. D., Satellite Communica- Seattle, Wash. Miller, L. E., Reliability of Sili- tion and Project Echo, Rotary Hansen, R. H., and DeBenedictis, con Transistors and Diodes. Ad- Club, Elizabeth, N. J. T., Studies of the Decomposi- visory Gp. on Electron Tubes Thurston, R. N., see Pfann, W. G. tion of Blowing Agents. I: A Conf. on Reliability of Semi- Trumbore, F. A., Problems in Method for Predicting Per- conductor Devices, N. Y. C. Semiconductor Crystal Growth, formance, Am. Chem. Soc., Moore, E. F., Machine Models of Phys. Chem. Sem., University Seton Hall University, South Self -Reproduction, RCA Labs., of Pittsburgh, Pittsburgh, Pa. Orange, N. J. Princeton, N. J. Tukey, J. W., The Future of Haugk, G., see Hoover, C. W. Mum, A. J., see Gianola, U. F. Data Analysis, Twenty- Fourth Hayes, J. S., Electron Beam Mi- Osgood, D. T., see Felder, H. H. Annual Meeting of Inst. of cronalyzer, New Tool for Elec- Pfann, W. G., New Ways of Math. Statistics, Seattle, Wash. tron Device Development, A.S. Using the Piezoresistance of Walsh, W. M., Magnetic Reso- T.M., Phil., Pa. Semiconductors to Measure nance in Nucleic Acid Samples, Hight, S. C., and Kreer, J. G., Stresses or Strains, University A. D. Little Co., Cambridge, Jr., Some Studies of Special of Illinois, Dept. of Mining & Mass. Orbital Configuration for Glob- Metallurgy, Urbana, Ill. Walters, K. R., Similarities of the al Common.ica tions, A.I.E.E./ Pfann, W. G., and Thurston, R. Human Nervous System and I.R.E. GLOBECOM V Conf., N.,Serniconcliictor Stress Trans- Our Communications Network, Chicago, Ill. ducers Utilizing the Trans- Princeton University, Depart- Hoover, C. W., and Haugk, G., verse and Shear Piezoresistance ment of Psychology, Princeton, The Flying Spot Store, Large Effects, Am. Phys. Soc., Wash., N. J. Capacity Memory Techniques D. C.; University of Pennsyl- Watson, B. J., see Flanagan, J. L. for Computing Systems Sym- vania, Metallurgy Depart. Col- White, D. L., Depletion Layer posium, Wash., D. C. loquium, Phil., Pa. Transducer-A New High Fre- Kaiser, W., The Optical Maser, Pierce, J. R., Research and Tech- quency Ultrasonic Transducer, Fifteenth Annual Frequency nology in the Space Age, Fourth I.R.E. National Cony., N. Y. C. Control Symposium, Atlantic Annual Alumni Seminar, Uni- Williams, W. H., Remarks on the City, N. J. versity of Kentucky, Lexington, Efficiency of Unbiased Estima- Kometani, T. Y., Determination Ky. tion with Auxiliary Variates, of Sodium Fluoride in Wood, Pierce, J. R., What Computers I.M.S. Meeting, Seattle, Wash. Am. Wood Preserver's Assoc., Can Do Better-And How, Wintringham, W. T., The Prin- Banff, Canada. M.I.T. Centennial Lecture Se- ciples of Color Television, Kreer, J. G., see Hight, S. C. ries, Cambridge, Mass. Thirtieth Annual Meeting of Leutritz, J., Jr., The Diinethyl- Pierce, J. R., What We Should Inter. Soc. Color Council, Roch- formamide (DMF) Extractables Do About Satellite Communica- ester, N. Y. in Steam -Conditioned Pine, Am. tion Now, First National Conf. Wright, J. P., Chemistry and Your Wood Preserver's Assoc., Banff, on Peaceful Uses of Space, Tul- Telephone, Rotary Club, Green- Canada. sa, Okla. ville, Tex. Lewis, W. D., The Impact of Com- Ruff, J. A., see Gianola, U. F. Yokelson, B. J., Electronic Switch- munication Technology on Com- Schawlow, A. L., Infrared and ing System, A.I.E.E., N. Y. C. puter Technology and Vice Optical Masers, Conf. on Atom- Zacharias, A., A Precise Method Versa, Information Processing ic Spectra, Argonne National for Measuring the Incremental & Computer Technology Sym- Lab., Argonne, Ill. Phase and Gain Variations of posium, Bendix Corp. Res. Labs. Schroeder, M. R., Further Prog- a Traveling -Wave Tube, 1961 Div., Southfield, Mich. ress with Colorless Artificial I.R.E. International Meeting, Linder, S. L., Improved Perform- Reverberation, Sixty -First N. Y. C.

August 1961 301

www.americanradiohistory.com PAPERS

Following is a list of the authors, titles, and places of publica- 1137 -1144, June, 1961. tion of recent papers published by members of the Laboratories. Guggenheim, H. J., The Prepara- tion of Single Crystals of Cer- tain Transition Metal Fuorides, Adler, R., see Engelbrecht, R. S. Burrus, C. A., and Trambarulo, J. Phys. Chem., 64 pp. 938 -939, Alburger, D. E., see Donovan, P. F. R. F., A Millimeter Wave Esaki July, 1960. Allen, F. G., Field Emission from Diode Amplifier, Proc. I.R.E., Haus, H. A., see Engelbrecht, Silicon and Germanium: Field 49, pp. 1075 -1076, June, 1961. R. S. Desorption and Surface Migra- Compton, V. B., see Bozorth, R. M. Hellman, M. Y., see Lundberg, tion, The Phys. & Chem. of Davis, D. D., see Bozorth, R. M. J. L. Solids, 19, pp. 87 -99, 1961. Donovan, P. F., "Recipes" for De- Ingram, S. B., Scientific and En- Aloisio, C. J., A Low Tempera - tector Fabrication "Paint -on gineering Manpower-The Im- ture Ultraviolet Cell, Rev. Sci. Particle Detectors (Recipe No. mediate Problem and Suggested Instr., 32, p. 452, Apr., 1961. 2) ", Semiconductor Nuclear Par- Solutions, Elec. Engg., 80, pp. Ashkin, A., A Low -Noise Micro- ticle Detectors National Acad- 341 -345, May, 1961. Re- wave Quadrupole Amplifier, emy of Sciences, National King, B. G., see Sharpe, G. E. 268- Proc. I.R.E., 49, pp. 1016 -1020, search Council, 871, pp. Kisliuk, P. P., The Reflection of June, 1961. 269, 1961. Slow Electrons from Tungsten D. E., Ashkin, A., see Gordon, E. I. Donovan, P. F., Alburger, Single Crystals, Clean and with Bateman, T., Mason, W. P., and Pixley, R. E., and Wilkinson, Adsorbed Monolayers, Phys. of N": Con- McSkimin, H. J., Third Order D. H., Beta -Decay Rev., 122, pp. 405 -411, Apr. 15, of and Parity in Elastic Moduli of Germanium, servation Spin 1961. 928 -936, 016, Phil. Mag., 6, pp. 171 -174, J. Appl. Phys., 32, pp. Kitsopoulos, S. C., and Kretzmer, Jan., 1961. May, 1961. E. R., Computer Simulation of P. A., Doleiden, F. H., see Fuller, C. S. Bozorth, R. M., Wolff, a Television Coding Scheme, L., see Van Uitert, L. G. Davis, D. D., Compton, V. B., Egerton, Proc. I.R.E., 49, pp. 1076 -1077, Engelbrecht, R. S., Adler, R., and Wernick, J. H., Ferromag- June, 1961. of Haus, H. A., Lebenbaum, M., netism in Dilute Solutions Klockow, D. H., Coupling Net- in. Phys. Rev., and Mumford, W. W., Elemen- Cobalt Palladium, work for an Esaki Diode Trans- 15, tary Considerations of Noise 122, pp. 1157 -1160, May mission Amplifier, Northeast- Performance, PGMT &T Na- 1961. ern University, B.L.E.P. Pro- Brown, W. L., Properties of Space tional Symposium Digest, 1, pp. gram, pp. 1 -45, Mar., 1961. Regions, Semiconductor 53 -57, May 15, 1961. Charge Kluver, J. W., Parametric Cou- Particle Detectors Pub- Frisch, H. L., see Lundberg, J. L. Nuclear pling Between the Transverse lication 871, National Academy Fuller, C. S., and Doleiden, F. H., Waves on o- and M- Type Sciences, National Research The Ionization Behavior of of Beams, J. Appl. Phys., 32, pp. Council, Wash., D. C., pp. 9 -18, Donors Formed from Oxygen in 1111 -1114, June, 1961. 1961. Germanium, The Phys. & Chem. Knab, E. D., Synchronic Index of Buck, T. M., Surface Effects on of Solids, 19, pp. 251 -260, May, Gear Trains, Prod. Engg., 32, Silicon Particle Detectors, Proc. 1961. pp. 35 -39, May 29, 1961. Asheville Conf. on Semiconduc- Gibson, W. M., Grown Oxide Edge Kitsopoulos, tor Nuclear Particle Detectors, Protection of p -n Junction Kretzmer, E. R., see 871, National Academy of Sci- Radiation Detectors, Proc. of S. C. ences, National Research Coun- Asheville Conf. on Semiconduc- Lebenbaum, M., see Engelbrecht, cil, pp. 111 -120, 1961. tor Radiation Detectors, Na- R. S. M. Y., Burrus, C. A., Gallium Arsenide tional Academy of Sciences, Na- Lundberg, J. L., Hellman, Esaki Diodes for High Fre- tional Research Council, 871, and Frisch, H. L., An Experi- quency Applications, J. Appl. pp. 232 -233, 1961. mental Study of Polymer Poly - Phys., 32, pp. 1031 -1036, June Golden, R. M., and Schroeder, dispersitrt by Viscometry, J. 1961. M. R., Discussion of "Combined Poly. Sci., 46, pp. 3 -17, Sept., Burrus, C. A., Gallium Anti- AM and FM for a One -Sided 1960. monide Esaki Diodes for High Spectrum," Proc. I.R.E., 49, p. Mason, W. P., see Bateman, T. Frequency Applications, Proc. 1094, June, 1961. McSkimin, H. J., see Bateman, T. I.R.E., 49, p. 1101, June, 1961. Gordon, E. I., and Ashkin, A., Meiboom, S., Loewenstein, A., and Burrus, C. A., High -Frequency Energy Interchange Between Fraenkel, G., Protona t ion in Silicon. Varactor Diodes, J. Cyclotron and Synchronous N-methylacetamide, J. Phys. Appl. Phys., 32, pp. 1166 -1167, Waves in Quadrupolar Pump Chem., 65, pp. 700 -702, Apr., June, 1961. Fields, J. Appl. Phys., 32, pp. 1961.

302 Bell Laboratories Record

www.americanradiohistory.com McSkimin, H. J., Notes and Ref- Piksis, A. H., see Schawlow, A. L. W. M., Jr., Williams, H. J., and erences for the Measurement Pixley, R. E., see Donovan, P. F. Wright, J. P., Ferromagnetic of Elastic Moduli by Means of Rich, A., see Shulman, R. G. Resonance in DNA Samples, Ultrasonic Waves, J. Acous. Rudsill, J. A., Jr., Servos Can Be Biochem.&Biophys. Res. Comm., Soc. Am., 33, p. 606, May, 1961. Designed Corner by Corner, V, pp. 52 -56, 1961. Miller, G. L., Diffused Junction Electronic Design, pp. 32 -35, Spencer, E. G., see Sansalone, F. J. Detectors, Semiconductor Nu- May 24, 1961. Sugano, S., see Schawlow, A. L. clear Particle Detectors Nation- Sansalone, F. J., and Spencer, Swanekamp, F. W., see Monforte, al Academy of Sciences, Na- E. G., Low Temperature Micro - F. R. tional Research Council, 871, wave Power Limiters, PGMTT, Trambarulo, R. F., see Burrus, pp. 19 -27, 1961. 9, pp. 272 -273, May, 1961. C. A. Monforte, F. R., Swanekamp, F. Schawlow, A. L., Piksis, A. H., Van Uitert, L. G., and Egerton, W., and Van Uitert, L. G., An and Sugano, S., Strain -Induced L., Bismuth Titanate -A Ferro - R. F. Technique for Pulling Effects on the Degenerate Spec- electric, J. Appl. Phys., 32, p. Oxide Crystals without Em- tral Line of Chromium in MgO 959, May, 1961. ploying a Crucible Susceptor, Crystals, Phys. Rev., 122, pp. Van Uitert, L. G., see Monforte, J. Appl. Phys., 32, p. 959, May, 1469 -1476, June 1, 1961. F. R. 1961. Walsh, W. M., Jr., see Shulman, Schroeder, M. R., see Golden, R. M. G. R. Mumford, W. W., see Engeibrecht, Sharpe, G. E., and King, B. G., Wernick, J. H., see Bozorth, R. M. R. S. Low Gain, Wide Band Esaki Wilkinson, D. H., see Donovan, Pierce, J. R., The Hazardous Diode Amplifiers, International P. F. Course of Solid Communications Sat- State Circuits Conf. Di- Williams, H. J., see Shulman, R. G. ellites, gest, pp. 98 -99, Bull. Atomic Sci., XVII, Feb., 1961. Wolff, P. A., see Bozorth, R. M. pp. 181 -185, May -June, 1961. Shulman, R. G., Rich, A., Walsh, Wright, J. P., see Shulman, R. G. PATENTS

Following is a list of the inve ntors, titles and patent members Kiltz, R. M. - Switching Appa- of patents recently issued to members of the Laboratories. ratus-2,988,672. Mallery, P. Magnetic Aamodt, - Control T.- Method of Forming Dunlap, K. S.- Crosspoint Switch- Circuits- 2,987,625. a Plug of High Melting Point ing Network Control System - Mallery, P. - Magnetic Memory Plastic Bonded to a Low Melt- 2,987,579. Arrangement-2,988,733. ing Point Plastic -2,989,784. Feinstein, J., see Collier, R. J. Malthaner, W. Cathode Fol- Abbott, A.- G. F., Jr., and Sumner, Fogarty, L. L. Magnetic Core lower Tube Circuit- 2,987,677. E. E.- Bistable Trigger - Circuit Counting Circuits- 2,989,647. Mason, W. P. - Stabilization of -2,987,628. Gianola, U. F., see Sharp, L. V. Quartz Crystal Frequency Berger, Con- U. S.-Cross-Modulation Grant, C. T. -Wave Filter trolling Elements -2,989,799. Measuring -2; System- 2,987,586. 990,525. Matthias, B. T. Ferromagnetic Callaway, W. B., and Deltuvia, A. Greiner, E. S. Processing of Material- 2,989,480. A., Jr. - Printing Apparatus - Boron Compact- 2,990,261. Mattke, C. F., see Hefele, J. R. 2,989,910. Hall, A. D., and Zarouni, Power, F. S. Manufacture of Collier, R. J., and Feinstein, R.- - J.- Tandem Echo Dry Electrolytic Devices Microwave Amplifier Suppressor Cir- -2,- - 2,988,- 2,990,457. 989,447. 669. cuits- Haring, H. E. Electric Batteries Restall, W. E., Jr., see Hershey, Courtney- Pratt, J. S. -High Speed H. J. Cine- Camera -U.K. 821,538. -2,988,587. Hefele, J. R., Seidel, H., and Weiss, J. A.-Mag- Courtney- Pratt, J. S. Piezore- and Mattke, C. F.- - Scanning Mechanism for Radio netically Controlled Wave Guide sistive Compensators - 2,990,- Switch-2,989,709. 529. Signaling Apparatus 2,987,- - Sharp, L. V., see Gianola, U. F. Cutler, C. C.- Traveling Wave 722. Hershey, Smith, L. M.- Transistor Pulse Tube -2,989,661. H. J., Jordan, H. G., and Generator-2,989,651. Deltuvia, A. A., Restall, W. E., Telephone Jr., see Callaway, Jr.- Spector, C. J., see Doucette, E. I. W. B. Dial Light -2,988,631. Sumner, E. E., see Abbott, G. F., Doherty, W. H. Hybrid Sideband Israel, J. 0.-Apparatus for Gen - Jr. erating Frequency Modulation. System Oscillations of Differ- Warner, R. M., Jr.- Semiconduc- -2,989,622. ent Frequencies- 2,987,680. tor Resistance Element- 2,989; Doucette, E. I., and Spector, C. J. Jordan, H. G., see Hershey, H. J. 713. -Semiconductor Capacitor -2,- Ketchledge, R. W. Position Con- Weiss, J. A., see Seidel, H. 989,650. trol System- 2,990,505. Zarouni, R., see Hall, A. D.

August 1961 303

www.americanradiohistory.com THE AUTHORS

Duncan H. Looney was born Scovil was a Nuffield Research in Muskogee, Oklahoma. He earn- Fellow at Oxford, and then re- ed the bachelor of science degree turned to the University of from Purdue University in 1948 British Columbia as Assistant and the Ph.D. from Massachu- Professor during 1952 -1955. setts Institute of Technology in In 1955 he joined Bell Tele- 1953. While working on his doc- phone Laboratories where, as a torate degree he was a research member of the Device Develop- assistant in the M.I.T. Research ment Department, he has done Laboratory of Electronics. extensive development work of Mr. Looney joined Bell Lab- masers. Recently, Mr. Scovil was oratories in 1953. During his first named head of the sub- depart- year with the Laboratories he ment responsible for solid -state was engaged in development of maser development. transistors and related semi-con- D. C. Hogg ductor devices. Since then he has John B. Bishop, author of "A worked on development of solid - search Department. He is a sen- New Surface -to -Air Data Com- state memory and logic devices. ior member of the IRE and in munication System" in this issue, 1960 was elected member of the is a native of Auburn, N. S. He re- Union Radio Scientifique Inter- ceived his B.Sc. from Acadia Uni- nationale (Commission 2) . He re- versity, Wolfville, N. S., in 1921, ceived a fellowship from the Na- tional Research Council of Cana- da in 1952. Mr. Hogg is co- author of the article on low -noise measurements appearing in this issue.

H. E. D. Scovil, co- author of the article on low -noise measurements, was born in Victoria, B.C.. Canada, and attended the Uni- versity of British Columbia, from D. H. Looney which he received the B.A. and 1949. He was recently named to head M.A. degrees in 1948 and he at Ox- a new sub -department in solid - Subsequently, studies B. Bishop state devices with emphasis on ford University, where he was J. awarded the D. Phil. degree in memory devices. and the Ph.D. in Physics from the 1951. During 1951 and 1952, Mr. Mr. Looney is a member of Cornell University in 1926. American Physical Society and After two years in design and the Institute of Radio Engineers. development at the Westinghouse He is the author of "Electronic Lamp Co., he joined the Labora- Memory Devices" in this issue. tories in 1928, working on the development of transmitters for David C. Hogg was born in commercial ship -to -shore and Vanguard, Saskatchewan, Cana- military communication systems, da. He received the B.Sc. degree until 1940. During the war years, from the University of Western he worked on the development of Ontario in 1949, and the M.Sc. various search and fire control degree and Ph.D. degree from radar projects. McGill University in 1950 and Immediately following the war, 1953 respectively. he spent two years developing FM Mr. Hogg joined Bell Labora- transmitters for broadcast appli- tories in 1953 in the Radio Re- H. E. D. Scovil cation. Since 1948, he has been in

304 Bell Laboratories Record

www.americanradiohistory.com AUTHORS CONTINUED; charge of a group responsible for ber of the AIEE, Phi Kappa Phi, the design and development of and Sigma Tau. weapons control and data communication equipment for the armed forces. A. F. Pomeroy was born at Buf- falo, Mr. Bishop N. Y. He graduated from is a member of the Brown University Institute of Radio in 1929 with a Engineers and B.S. Sigma Xi. degree in Engineering and joined Bell Laboratories that year. His early work was devoted to W. B. Callaway, a native of developing test and measurement Florence, Alabama, received his equipment and the study of trans- B.S. in E.E. from Alabama Poly- mission lines. His work with man- technic Institute in 1938. After a ufacturers has resulted year in the de- with the Varityper Corpora- velopment of waveguides tion, with he joined the Laboratories J. W. Osmun improved transmission qualities. in 1940, working in various trans- Since 1956 he has been mission, concerned switching, and signaling engaged in systems with the development development engineering of ferrite projects. During studies, particularly in the appli- devices such as microwave attenu- World II, War Mr. Callaway cation of computers and ators, isolaters and switches. Early spent three years data in the U. S. processing techniques to the meas- last year, Mr. Pomeroy trans- Naval Reserve, developing equip- urement and analysis of telephone ferred to Western Electric ment and tactics at the for antisubma- traffic. He is the author of "The Merrimac Valley Works. A mem- rine warfare, as well as in opera- Art of Counting Calls" in this ber of the IRE, he is the author tional flying in both the South issue. of "The Ferrite Isolator: A Atlantic and Pacific New theatres. Kind of Pad for TJ Radio -Relay" Most recently, he has been J. W. Osmun was born and in this issue. raised in Sparks, Nevada and now lives in Scotch Plains, N. J. From 19-13 to 1947 he served with the U. S. Army in the South Pa- cific as a parachutist. Mr. Osmun received a B.S.E.E. from Univer- sity of Nevada in 1953, and that year he joined the Laboratories and graduated from the CDT program in 1956. A member of the Power Development Department, he has specialized in ringing power plants and transistorized dc-dc power converters. Author of "A Statistical Ammeter" in W. B. Callaway this issue, Mr. Osmun is a mem- A. F. Pomeroy

www.americanradiohistory.com What was Bell Telephone Laboratories doing ON FRIDAY, JUNE 30, 1961?

the development of a It was exploring the communications pos- It was preparing an experiment in world- It was completing route" Long Distance micro- sibilities of the gaseous optical maser - wide communications using "active" satel- new "heavy system capable of handling over a device which generates continuous co- lites powered by the solar battery, a Bell wave two -way conversations at once. herent infrared radiation in a narrow beam. Laboratories invention. 11,000

with an electronic It was developing an anti -missile defense It was demonstrating the potential'i'ties of It was experimenting at Morris, Ill., which is capable system designed to detect, track, intercept the superconducting compound of niobium central office of providing a wide range of new telephone and destroy an enemy ICBM - in a matter and tin for generating, with little power, of minutes. magnetic fields of great strength. services.

repeaters or It was continuing its endless search for It was perfecting the card dialer which It was developing improved new knowledge under the leadership of permits, through insertion of a punched "amplifiers" to increase greatly the capacity economy of undersea telephone cable scientists and engineers with world -wide card into a slot, automatic dialing of fre- and reputations in their chosen fields. quently used numbers. systems.

Bell Laboratories scientists and engineers work with every art and science that can benefit communications. Their inquiries range from the ocean floor to outer space, from atomic physics to the design of new telephone sets, from the tiny transistor to massive transcontinental radio systems. The goal is constant - ever -improving Bell System communications services. BELL TELEPHONE LABORATORIES World center of communications research and development