Showing papers in "Bell System Technical Journal in 1967"

The si-sio, interface – electrical properties as determined by the metal-insulator-silicon conductance technique

Abstract: Measurements of the equivalent parallel conductance of metal-insulator-semiconductor (MIS) capacitors are shown to give more detailed and accurate information about interface states than capacitance measurements. Experimental techniques and methods of analysis are described. From the results of the conductance technique, a realistic characterization of the Si–SiO 2 interface is developed. Salient features are: A continuum of states is found across the band gap of the silicon. Capture cross sections for holes and electrons are independent of energy over large portions of the band gap. The surface potential is subject to statistical fluctuations arising from various sources. The dominant contribution in the samples measured arises from a random distribution of surface charge. The fluctuating surface potential causes a dispersion of interface state time constants in the depletion region. In the weak inversion region the dispersion is eliminated by interaction between interface states and the minority carrier band. A single time constant results. From the experimentally established facts, equivalent circuits accurately describing the measurements are constructed.

A floating gate and its application to memory devices

Abstract: A structure has been proposed and fabricated in which semipermanent charge storage is possible. A floating gate is placed a small distance from an electron source. When an appropriately high field is applied through an outer gate, the floating gate charges up. The charges are stored even after the removal of the charging field due to much lower back transport probability. Stored-charge density of the order of 1012/cm2 has been achieved and detected by a structure similar to an metal-insulator-semiconductor (MIS) field effect transistor. Such a device functions as a bistable memory with nondestructive read-out features. The memory holding time observed was longer than one hour. These preliminary results are in fair agreement with a simple analysis.

Factoring polynomials over finite fields

Abstract: We present here an algorithm for factoring a given polynomial over GF(q) into powers of irreducible polynomials. The method reduces the factorization of a polynomial of degree m over GF(q) to the solution of about m(q − 1)/q linear equations in as many unknowns over GF(q).

An adaptive echo canceller

Abstract: A novel method is presented for echo-cancellation in long distance telephone connections. In contrast with conventional echo suppressors, the device described achieves echo-cancellation without interrupting the return path. A replica of the echo is synthesized and subtracted from the return signal. The replica is synthesized by means of a filter which, under the control of a feedback loop, adapts to the transmission characteristic of the echo path and tracks variations of the path that may occur during a conversation. The adaptive control loop is described by a set of simultaneous, nonlinear, first-order differential equations. It is shown that under ideal conditions, the echo converges to zero. Estimates of the rate of convergence are obtained. Effects of noise are discussed. The results of computer simulations of various alternative configurations of the system are described.

Properties and device applications of magnetic domains in orthoferrites

Abstract: It has been shown that isolated magnetic domains in thin platelets (≍2 mils thick) of orthoferrites can be manipulated to perform memory, logic, and transmission functions. The purpose of this paper is to discuss the properties of orthoferrites that make them suitable for magnetic device applications and consider magnetostatic problems relevant to domain structures found to be useful. Included is a brief indication of how memory, logic, and transmission can be accomplished; however, the details will be reserved for a later paper. The stability conditions of a cylindrical domain are discussed in detail and data is reported to support the conclusions. Of particular interest are the sizes of cylindrical domains available in the various orthoferrites. Such data has been taken on five of the fourteen possible orthoferrites and it is found that the thulium orthoferrite, TmFeO 3 , gives the smallest stable domain diameter (2.3 mils) and LuFeO 3 the largest. The stability results lead to a direct method for obtaining σ W , the domain wall energy density. For TmFeO 3 , as an example, σ W = 2.8 ergs/cm2. It is concluded that the orthoferrites are well suited indeed for device applications. Experimentally, 3 mil diameter domains have been manipulated and there is every reason to believe that operation of sub-mil domains will soon be realized.

Transistor distortion analysis using volterra series representation

Abstract: Intermodulation distortion due to nonlinear elements in transistors is analyzed using Volterra series representation. It is shown that this technique is well suited for the analysis of transistor distortion where the nonlinearities are small but frequency dependent. An ac transistor model incorporating four nonlinearities is briefly described. The nonlinear nodal equations of the model are successively solved by expressing nodal voltages in terms of the Volterra series expansion of the input voltage. Based on this analysis, a digital computer program has been developed which computes the second and the third harmonic distortion for a given set of input frequencies and transistor parameters. The results compare favorably with measured values. This method also enables the derivation of closed form ac expressions for a simplified model; these expressions show the dependence of distortion on frequency, load and source impedances, bias currents and voltages, and the parameters of the transistor. The technique is also extended to cascaded transistors, and simplified expressions for the overall distortion in terms of the distortion and gain of individual transistors are derived. Finally, a few pertinent practical applications are discussed.

Minimum cost communication networks

Abstract: Cities A 1 , …, A n in the plane are to be interconnected by two-way communication channels. N(i, j) channels are to go between A i and A i . One could install the N(i, j) channels along a straight line, for every pair i, j. However it is usually possible to save money by rerouting channels over longer paths in order to group channels together. In this way, large numbers of channels share such preliminary expenses as real estate, surveying, and trench digging. The geometry of the least expensive network will depend on the numbers of channels N(i, j) and on the function f(N) which represents the cost per mile of installing N channels along a common route. If the preliminary expenses are the only expenses then f(N) is a constant, independent of N. In that case the best network is obtained by routing channels along lines of the “Steiner minimal tree”, a graph which has been studied extensively and which can be constructed by ruler and compass. In part, this paper generalizes Steiner minimal trees for the case of an arbitrary function f(N). One again obtains a ruler and compass construction for a minimizing tree, which is likely to provide a best or good solution when preliminary costs are a significant part of the total cost. However the minimizing tree need not be the best solution in general because further cost reductions may now be possible by using graphs which have cycles. Other properties of Steiner minimal trees generalize only part way, and some examples illustrate the new complications. The remainder of the paper considers functions f(N) with special properties. A convexity property $f(N+2)-2F(N+1)+f(N)\leqq 0, N\ =\ 1,2,\ldots$ ensures that there is a minimizing solution in which all N(i, j) channels between A i and A i take the same path (no split routing). If f(N) is a linear function (f(N) = a + bN), one can obtain simple bounds on the minimum cost. The lower bound is fairly accurate.

Some simple self-synchronizing digital data scramblers

Abstract: Two types of self-synchronizing digital data scramblers and descramblers are introduced and examined. The descramblers recover synchronization quickly after the insertion or deletion of channel bits, and they are relatively insensitive to channel errors. The scramblers act to increase the period of periodic data sequences, and the periodic channel sequences produced have approximately half as many transitions in one period as there are bits in a period. These circuits find application in common carrier systems where short-period data sequences produce high-level tones in the transmission band and, as a consequence, interchannel interference. And they have application when receiver clocks derive synchronization from transitions in the channel signal. A number of variations and modifications of the scramblers which affect their cost and size are considered. The scramblers and descramblers are similar in construction and consist of linear sequential filters with either feed-forward or feedback paths, counters, storage elements and peripheral logic. The counters, storage elements and peripheral logic monitor the channel sequence but react infrequently so that the scramblers and descramblers behave principally as linear sequential filters.

Nonlinear optical coefficients

Abstract: We consider, from a number of different viewpoints, the tensor coefficients which describe second harmonic generation, optical rectification, and the Pockels or linear electro-optic effect in acentric crystals. Stationary perturbation theory is used to calculate the low-frequency limit of the intrinsic electronic nonlinearity neglecting all effects due to local fields or lattice polarization. Solid methane is used as an example and the result used to estimate the coefficient in hexamethylene tetramine. The calculated result is within a factor of 2 of the experimental figure. The method is susceptible to further refinement and, since it requires only a knowledge of ground state wave functions, and is essentially very simple, it appears to offer a useful approach to the calculation of the coefficients. The classical anharmonic oscillator model is briefly covered and the model is related to a quantal treatment. We find that the anharmonic potential used in the model is directly related to the actual crystalline potential. It can also be related to the charge distribution in the electronic ground state. Local field corrections and the effects of lattice polarization are presented. These alter the nonlinear properties in a simple and obvious way, but one which has been misunderstood in some of the literature. Our results form a theoretical background to Miller's empirical rule relating the nonlinear coefficients to the linear susceptibilities. An extensive table of Miller-reduced tensor coefficients collated from the published literature is presented. Finally, we draw together some of the threads of the previous sections. An appendix deals with the vexing question of definitions.

Floating-point-roundoff accumulation in digital-filter realizations

Abstract: In this paper, several results are presented concerning the effects of roundoff in the floating-point realization of a general discrete filter governed ideally by a stable difference equation of the form $w_{n}\ =\ \sum_{k=0}^{M}\ b_{k}_{n-k}\ -\ \sum_{k=1}^{N}\ a_{k}w_{n-k}, \quad n \geqq N \qquad \eqno{\hbox{(1)}}$ in which {w n } and {x n } are output and input sequences, respectively. In particular, for a large class of filters it is proved that there is a function f(K) with f(K) → 0 as K → ∞ and a constant c, both dependent on the b k , the a k , the order in which the products on the right side of (1) are summed in the machine, and t, the number of bits allotted to the mantissa, such that $\langle e \rangle _{k} \leqq c \langle y \rangle_{k}+f(K)$ for all K ≧ N, in which, with {y n } the computed output sequence of the realized filter, $\langle y \rangle_{k}=\left({1 \over k+1} \sum_{n=0}^{k}\vert y_{n}\vert^{2}\right)^{{1 \over 2}}$ and $\langle e \rangle_{k}=\left({1 \over k+1} \sum_{n=0}^{k}\vert w_{n}- y_{n}\vert^{2}\right)^{{1 \over 2}}$ Bounds on f(K) and c are given that are not difficult to evaluate, and which, in many realistic cases, are informative. For example, for the second-order bandpass filter: $w_{n}=x_{n}a_{1}w_{n-1}\ - \ a_{2}w_{n-2}, \quad n \geqq 2 \qquad \eqno{\hbox{(2)}}$ with a 1 and a 2 chosen so that its poles are at approximately ± 45° and at distance approximately (but not less than) 0.001 from the unit circle, we find that c, an upper bound on the “asymptotic output error-to-signal ratio”, is not greater than 0.58 × 10−4 assuming that t = 27, that the terms on the right side of (2) are summed in the machine in the order indicated from right to left), and that the x_{n} in (2) are machine numbers. I f the x_{n} are not machine numbers, and hence must be quantized before processing, then c ≦ 0.76 × 10−4. In addition to error bounds, an inequality is derived which, if satisfied, rules out certain types of generally undesimble behavior such as self-sustained output limit cycles due to roundoff effects. This inequality is satisfied for the example described above.

A generalized nyquist criterion and an optimum linear receiver for a pulse modulation system

Abstract: A pulse modulation system is modeled with M waveforms {s m (t)}M 1 , each of which is amplitude scaled and simultaneously transmitted over a single physical channel. An infinite pulse train is assumed with signal interval T, which is determined by bandwidth consideration of the channel. We restrict the receiver to be linear with M outputs, one for each signal waveform. At a high signal-to-noise ratio the main sources of interference at the input to the receiver are the intersymbol interference and crosstalk; by crosstalk we mean the interference between the different waveforms. It is desirable, therefore, for the receiver to eliminate both types of interference and to minimize the remaining error due to additive noise in the channel. This constraint on the intersymbol interference and crosstalk is defined as the generalized Nyquist criterion. The receiver which accomplishes the above is determined for a mean square error criterion. Finally, some examples are presented which demonstrate the ease with which the generalized Nyquist criterion can be used to design waveforms without intersymbol interference or crosstalk.

Noise-like structure in the image of diffusely reflecting objects in coherent illumination

Abstract: Holographic and other imaging systems utilizing coherent light introduce a speckled or noise-like pattern in the image of a diffuse object which severely degrades image quality It is desirable to understand this effect quantitatively Intelligent design in many cases requires knowledge of the mean-square value, spatial power spectral density, and autocorrelation junction of the noise-like fluctuations These quantities have been determined for the image of a uniform diffuse object Major results are: (i) The mean-square value of the fluctuation in the image intensity is equal to the square of the mean intensity (ii) One can decrease the relative magnitude of the noise-like fluctuations at the cost of a corresponding increase in the aperture required of the optical system (or hologram) over that required to resolve the desired image in a spatial frequency sense In a holographic facsimile or TV system, this calls for a corresponding increase in electrical bandwidth (iii) The improvement in (ii) is not possible for direct viewing with the human eye, since the resolution of a healthy eye is known to be limited by diffraction at the iris

A normal limit theorem for power sums of independent random variables

Abstract: Suppose that $p_{n}\ = \ 10 {\rm log}_{10}\ [10^{x_{1}/10} + \ldots + 10^{x_{n}/10}]$,$ where {X n } is a sequence of independent random variables. The main result of this paper shows that under very general conditions on the sequence {X n }, the power sums P n will be asymptotically normally distributed. This result supports a commonly used normal approximation, and shows why many physical quantities obtained by power addition of random variables tend to be normally distributed in dB.

An approach to a unified theory of automata

Abstract: A model of an automaton, called a balloon automaton is proposed, It consists of a finite control, which may be deterministic or nondeterministic, an input tape which may be one way or two way, and an abstract, infinite memory, called the balloon, which can enter any of a countable number of states. There is assumed to be a recursive function which manipulates the state of the balloon, and another which passes a finite amount of information from the balloon to the finite control. A subset of the balloon automata is considered a closed class if it obeys two very simple closure properties. Certain closed classes recognize exactly the languages recognized by such familiar automata as the pushdown automaton or stack automaton. Unfortunately, no closed class recognizes the sets accepted by linear bounded automata or the time and tape complexity classes of Turing machines. It is shown that many of the usual closure properties of languages accepted by the pushdown automaton, stack automaton, etc., hold for an arbitrary closed class of balloon automata. For example, the languages accepted by a closed class of one-way, nondeterministic balloon automata are closed under concatenation. Of special interest is the fact that a closed class of two-way deterministic balloon automata is closed under inverse g.s.m. mappings. This fact is not obvious, and was not known for all of the types of automata which form closed classes of balloon automata. It should be emphasized that the purpose of this paper is not to propose another “model of a computer.” Rather, we are proposing a method of proving the standard theorems about existing and future models. Hopefully, when a model is proposed in the future, one will simply show it equivalent to a closed class of balloon automata, and have many of the closure properties automatically proven.

Random packings and coverings of the unit n-sphere

Abstract: It is well known that the quantity M v (n, θ), the maximum number of nonoverlapping spherical caps of half angle θ (a “packing”) which can be placed on the surface of a unit sphere in Euclidean n-space is not less than exp [– n log sin 2θ + o(n)] (θ c (n, θ), the minimum number of caps of half angle θ required to cover the unit Euclidean n-sphere. We show that M c (n, θ) = exp [–n log sin θ + o(n)]. The central part of the proof is also a random coding argument which asserts that if a set roughly exp (–n log sin θ) caps is chosen at random, that on the average only a very small fraction of the surface of the n-sphere will remain uncovered (when n is large).

Surface effects of radiation on semiconductor devices

Abstract: A brief review of surface physics is given as background for the subsequent discussion on the role of surfaces in the behavior of semiconductor devices. The effects of channels and surface generation-recombination on p-n junctions and transistor characteristics are discussed. The observed effects of ionizing radiation on nonpassivated, gas-filled transistors are interpreted in terms of a model in which ions formed in the gas ambient deposit charge on the device surface. The resultant surface charge buildup creates channels on the device surface which cause a decrease in h FE and increase in I CBO . Saturation, recovery, and the effects of dose rate and bias are also discussed. Degradation of planar passivated transistors and other devices employing SiO 2 layers due to radiation is similar to that observed for nonpassivated devices. Surface charge buildup affects the device surface and leads to degradation. The bulk of experimental evidence points to accumulation of positive charge at the SiO 2 –Si interface as the cause of degradation. Several possible means of charge buildup at the interface are discussed. However, the process responsible has not, as yet, been identified. The direction of future experiments is discussed, particularly of those experiments which may yield information about the part played by radiation in positive charge accumulation at the SiO 2 –Si interface.

An automatic equalizer for general-purpose communication channels

Abstract: The restriction imposed by linear distortion on the flow of information in a communication channel is well known. In the past, the effects of this distortion have been alleviated through the use of manually adjusted equalizing or compensating networks. The adjustment of these networks is too cumbersome a process for the user of a switched communication service to perform each time a new connection is established. Therefore, in present switched networks, control of linear distortion is imposed only on the individual links. Variation between links and variation of the number of links in tandem result in channels with distributed performance. Lower distortion can be achieved by equalizing the overall connection. Recent developments have made automatic linear distortion removal (equalization) practical for synchronous data communication systems. Here an implementation is described wherein these techniques have been generalized so that automatic equalization can be provided for a communication channel independent of the signal format used in that channel. For a number of applications the speed of automatic equalization makes efficient end-to-end equalization practical in a switched network. The implementation described affords automatic minimization of the discrepancy between a specified response and the actual response of a linear transmission medium. Thus, on the one hand, it permits the automatic reduction of transmission defects such as signal dispersion and echoes, and, on the other hand, it permits the mechanized synthesis of filters with specified transfer functions. This paper reviews the general aspects of automatic equalization, describes an implementation of a general purpose automatic equalizer, discusses the theoretical performance of such an equalizer as determined from computer simulations, and lastly presents results for the equalization of real channels using the implementation described.

Restoration of photographs blurred by image motion

Abstract: The blurring of photographs by image motion during exposure is studied by means of a simple model. Conditions under which it is possible to recover the unblurred image are determined and some methods of restoration are described.

B.S.T.J. BRIEFS: A camera tube with a silicon diode array target

Abstract: A variety of electronic cameras have been developed for television systems.1 Among these the vidicon2 and the Plumbicon3 have the inherent advantages of high sensitivity, small size, and simple mechanical construction. The operating principles of the vidicon and the Plumbicon are quite similar since they both utilize a thin photoconductive layer to convert the optical image to a stored charge pattern which is periodically scanned and erased by an electron beam. Erasing the charge pattern creates the video signal. However, there is a distinct difference in overall device performance since the photoconducting target in the Plumbicon (PbO) is deposited in a manner to form a single, large area, graded p-n junction, each layer having high resistivity. In the vidicon, the evaporated layers of Sb 2 S 3 forming the target behave like a semi-insulating photoconductor.

Timing recovery for synchronous binary data transmission

Abstract: This paper analyzes different methods of adjusting the sampling time for detecting synchronous binary data, based on properties of the random data signal itself The static error and the variance of the jitter of the resultant sampling instant are calculated where the effects of frequency offset, additive noise, signal overlap, and jitter of the reference source are included The threshold crossing timing recovery system adjusts the sampling time in response to the times at which the data signal crosses the amplitude threshold The sampled-derivative system uses the time derivative of the signal at the sampling time to adjust sampling phase It is shown that both systems lead to approximately the same amount of jitter in the presence of noise and signal overlap for a given bandwidth of the control loop An improved timing recovery system is presented which is constructed by adding correction signals to the sampled-derivative system This system accounts for intersymbol interference in a manner that tends to set the sampling time at the point of maximum eye opening, where the error probability is minimum for the most adverse message sequence

Low-resolution TV: Subjective effects of frame repetition and picture replenishment

Abstract: Using the experimental television facility described in a companion paper, frame repeating and point-by-point selective replenishment of picture elements have been accomplished in real time. On the basis of initial experiments, using the head-and-shoulder view of a person as the picture source, such as is likely to be encountered in a visual communication system, the following tentative conclusions have been reached: (i) The motion rendition with a 15 new pictures/second frame repeating system, while not flawless, is reasonably good. (ii) Selectively replenishing one-quarter of the picture points per frame gives a better continuity of motion but results in objectionable patterns. (iii) Picture quality greatly depends on the pattern of picture replenishment. Of the five replenishment patterns tested, two result in pictures which are significantly better than the other three. (iv) In informal viewings, opinion has been so divided that no preference has been established between simple 15 new pictures/second frame repeating and the more satisfying schemes for picture replenishment. (v) The frame repeating and replenishment systems produce gross impairment during zooming and panning; consequently, these systems in their present form are unlikely to be useful for broadcast television. The impairments observed in these systems are subjective and not yet predictable. This emphasizes the importance of subjective testing of systems in real time.

Temperature dependence of inversion–layer frequency response in silicon

Abstract: Conductance-voltage and capacitance-voltage curves of metal-oxide semiconductor (MOS) capacitors on n-type silicon were investigated in the temperature range between room temperature and 200°C Plots of the inversion-layer conductance versus reciprocal temperature show a sequence of two activation energies: one corresponding to the temperature dependence of the intrinsic carrier density n i , the other to that of n2 i The low-temperature range is characterized by recombination-generation in the space-charge region, the high-temperature range by diffusion current from the bulk The technique permits measurement of bulk lifetime for the two regimes and determination of room temperature cutoff frequency for the channel

Statistical analysis of the level crossings and duration of fades of the signal from an energy density mobile radio antenna

Abstract: A theoretical analysis of signal fading using an energy density antenna is developed and compared with that from an isotropic antenna. The energy density antenna provides a signal proportional to the energy density of the mobile radio field. The number of crossings that the signal makes of a given signal level and the average duration of fades below a given signal level have been derived theoretically for these two cases using a simple statistical model. Comparing the number of level crossings of the electric field with that of the energy density, it is shown that the energy density fades less frequently than the electric field by at least a factor of two. The average duration of fades of the electric field is greater than that of the energy density only for lower signal levels. These results are in reasonable agreement with experimental measurements.

The enumeration of information symbols in BCH codes

Abstract: This paper presents certain formulas for I(q, n, d), the number of information symbols in the q-ary Bose-Chaudhuri-Hocquenghem code of block length n = qm − 1 and designed distance d. By appropriate manipulations on the m-digit q-ary representation of d, we derive a simple linear recurrence for a sequence whose mth term is the number of information symbols in the BCH code. In addition to an exact solution of all finite cases, we obtain exact asymptotic results, as n and d go to infinity while their ratio n/d remains fixed. In this limit, the number of information symbols increases as n'. Specifically, we show that for fixed u, 0 ≦ u ≦ 1, $\lim\limits_{m \rightarrow \infty}\ q^{-ms} I(q,q^{m}\ -\ 1, uq^{m})=1$ where s is a singular function of u. The function s(u) is continuous and monotonic nonincreasing; it has derivative zero almost everywhere. Yet s(0) = 1 and s(1) = 0. For q = 2, s(u) is plotted in Fig. 1.

Some properties of power sums of truncated normal random variables

Abstract: The power sum of P n n components X 1 , X 2 , …, X n is defined by the relation $p_{n}\ = \ 10 {\rm log}_{10}\ [10^{X_{1}/10} + \ldots + 10^{X_{n}/10}]$ The distributions of such power sums are studied both analytically and by Monte Carlo simulation techniques for the case where the components are independent, identically distributed, truncated normal random variables. Results are given in terms of distributions and moments of P n . The number of components varies from 2 to 256, and the standard deviation of the component variables before truncation ranges from 1 to 10 dB. The dependence of the results on the choice of truncation point is also investigated.

The charge-control concept in the form of equivalent circuits, representing a link between the classic large signal diode and transistor models

Abstract: It is shown in this paper that the charge-control concept can be conceived as a special form of the Linvill model for semiconductors. Instead of mathematical tools, charge-control models become equivalent circuits amenable to ordinary network analysis techniques. In the simplest form, the charge-control equivalent circuit for the junction transistor is fully equivalent to the Linvill and the Beaufoy-Sparkes model. For all practical purposes, it is also equivalent to the Ebers-Moll model. The charge-control junction transistor equivalent circuit combines those features of the other models that are important for electrical engineering applications. It also permits the conversion between the three basic types of models. Because of its close relationship to the physical processes governing a device, it can readily be extended to higher-order phenomena. This is usually done by expressing a Linvill-type lumped model in terms of charge parameters. The charge-control equivalent circuit can be useful for modeling a variety of semiconductor devices.

Design considerations for a semipermanent optical memory

Abstract: The potential of high-speed optical memories using electro-optic or acousto-optic light deflection for address selection is examined. It is shown that for such memories the total memory capacity decreases as the third power of the addressing rate and that capacities in excess of 108 bits are feasible with a random access rate of 106 addresses/sec. A specific semipermanent memory design is then described which uses a laser light source, an acoustic xy light deflector and an array of 104 holograms as information storage elements. Each storage element contains 104 bits which appear as a pattern on a semiconductor read-out matrix when the storage element is illuminated through the xy deflector. Accordingly, the system has a total capacity of 108 bits with an access time of less than 10 μsec.

Slope overload noise in differential pulse code modulation systems

Abstract: In differential pulse code modulation (DPCM) systems, often referred to as predictive quantizing systems, the quantizing noise manifests itself in two forms, granular noise and slope overload noise. The study of overload noise in DPCM may be abstracted to the following stochastic processes problem. Let the input to the system be a Gaussian stochastic process {x(t)} with a bandlimited (0, ∫ 0 ) spectrum F(f). Denote the output of the system by y(t). Most of the time y(t) is equal to x(t). During time intervals of this kind, the absolute value of the derivative x'(t) = dx(t)/dt is less than a given positive constant x' 0 . (In a DPCM system, x' 0 = kf. where k is the maximum level of the quantizer and ∫ s is the sampling frequency.) There are time intervals, I i (t(i) 0 , t(i) 1 ) (i = 0, ± 1, ±2, …), for which y(t) ≠ x(t). These time intervals begin at time instants t(i) 0 such that | x'(t(i) 0 ) | increases through the value x' 0 . For t ∊ I i , y(t) = x(t(i) 0 ) + (t − t(i) 0 )x' 0 . The interval ends at t(i) 1 , when x(t) and y(t) become equal again. The overload noise in the DPCM system is defined to be n(t) = x(t) − y(t). The problem is to study the random process {n(t)}. In the present paper, we will give an upper bound to the average noise power $\langle n^{2}(t)\rangle_{av}$ which at the same time is a very good approximation to the noise power itself. Two previous attempts have been made to find $\langle n^{2}(t)\rangle_{av}$ . One, due to Rice and O'Neal, involves an approximation valid only for very large x' 0 . Another approach to the problem, due to Zetterberg, includes an ingenious way of avoiding the determination of t(i) 1 . A new approach is given here that combines the best features of the two methods. The present result is a better approximation for slope overload noise than has been previously obtained. The result differs from previous results but is asymptotically equal to that given by Rice and O'Neal for x' 0 → ∞. In the region where overload noise is important, the present result is in very good agreement with computer simulation and experiment. The technique used could be applied for the determination of other statistical characteristics of the error random process.

Subjective evaluation of transmission delay in telephone conversations

Abstract: An earlier experiment by Riesz and Klemmer on the effect of pure-transmission delay upon natural telephone conversations was extended in a test with more than double the time period and number of calls. The previous finding of little or no adverse reaction to round-trip pure delays of 600 and 1200 msec alone was confirmed. The previous finding of a large increase in dissatisfaction with both of these delays following exposure to 2400 msec was not obtained. Exposure to delays of 2400 msec led to no dissatisfaction with later calls at 600 msec, but some rejections at 1200 msec did occur. There is no contradiction of other results on normal telephone circuits with 2-wire terminations (and related echo sources, paths, and suppressors) wherein customer dissatisfaction is greater with 600 msec delay than with the much shorter delay of a normal long-distance circuit.

A quantitative theory of 1/f type noise due to interface states in thermally oxidized silicon

Abstract: A quantitative theory of 1/f type noise is derived from the distribution of trapping times for charges in interface states. The distribution of trapping times has been recently explained quantitatively by means of a random distribution of surface potential caused by a random distribution over the plane of the interface of fixed charges located in the oxide. This model, which agrees with the interface state time constant dispersion measured by the MIS conductance technique, leads to a noise spectrum which is independent of frequency at very low frequencies, tends towards a 1/f2 dependence at high frequencies, and has an extended 1/f frequency dependence at intermediate frequencies. The mechanism for time constant dispersion is independent of temperature and silicon resistivity; it depends only on the majority carrier density at the silicon surface, the interface state density, and the density of fixed oxide charges. The dependence of open circuit mean square noise voltage on these parameters and frequency are illustrated for an MOS capacitor.