Comments, with reply, on `Random telegraph noise ofdeep-submicrometer MOSFETs' by K.K. Hung et al

For original paper see ibid., vol.11, no.2, p.190 (1990). The commenter discusses several points in the above-titled correspondence that appear questionable. One of these points concerns the assertion according to which ϶N/϶N_t≃-1 when N>10¹¹/cm². It is pointed out that this is true for non-VLSI types of devices, but it is not acceptable for ULSI devices with 86-Å-thick gate oxide and 5×10¹⁷/cm³ channel coping. A second point of criticism concerns the necessity of taking into account the mobility fluctuations induced by that of the trapped carrier number. It is suggested that before embodying this effect, it would be better to account for the gate voltage dependence of the effective mobility in strong inversion due to the vertical electric field. Such a treatment is outlined. The authors show, in the first case, that the error caused by the approximation is negligible and, in the second case, that they have taken into account the gate voltage dependence of the effective mobility in the analysis of the random telegraph noise data, although not explicitly     

Comments, with reply, on `Impact ionization in GaAs MESFETs' by K.Hui, et al

In the above-named work (see ibid., vol.11, p.113-15, March 1990), Hui et al. proposed a method to measure impact ionization current in GaAs MESFETs and evaluated the impact ionization coefficient α_n in GaAs. For electric fields greater than approximately 1.5×10⁵ V-cm^-1, α_n can be fitted to the equation α_n=4.0×10 ⁶×exp (-2.3×10⁶/E). In the present work, the commenters performed careful measurements of gate current I_g in GaAs MESFET devices similar to those used by Hui et al., and they show that the ionization coefficient still fits the above equation down to α_n=10^-4 cm^-1 . These results extend the previous data by three orders of magnitude. In a reply, the original authors affirm that the commenters have significantly improved the accuracy of the data previously presented     

Comments, with reply, on 'Another algorithm for converting a low-pass function into bandpass form' by K. Metzger

The commenter points out that the recursive algorithm described by K. Metzger in the above-titled letter (ibid., vol.76, no.5, p.634-5, May 1988) was published by him in the September 1980 issue of the IEEE Circuits and Systems Magaz (vol.2, no.3, p.3-5). He notes that the recursive technique is very useful not only for the low-pass to bandpass transformation, but also for performing the bilinear transformation used in digital filter design. The author replies that he was unaware of the earlier paper and thanks the commenter for calling it to his attention.<>     

Comments, with reply, on `AlGaAs/GaAs HBT for high-temperatureapplication' by K. Fricke et al

The commenter discusses several aspects of the above-titled work by K. Fricke et al. (ibid., vol. 39, pp. 1977-1981, Sept. 1992). These concern the physics of the saturation in an HBT and the collector leakage current. In their reply, to demonstrate the second effect, Fricke et al. report that the collector leakage current at different ambient temperatures was measured in common base configuration. They discuss the influence of the hole leakage current on the base current     

Comments, with reply, on `Operation of SOI CMOS devices atliquid-nitrogen temperature' by K.K. Young and B.-Y. Tsaur

The commenters discuss the increase factor of maximum electron mobility when NMOS SOI devices cool down to liquid-nitrogen temperature (LNT) reported by the authors of the above-mentioned paper (ibid., vol.11, p.126-8, Mar. 1990). The commenters point out that the maximum increase in mobility of only a factor of two over that of room temperature is significantly lower than the increase of four to five times that is generally accepted, and they report experimental results supporting the higher value. The authors in replying have recalculated the effective mobility based on the linear-current equation, and state that the LNT operation of the SOI devices might suffer from poor SiO₂-Si interface properties which result in less electron mobility increase than in bulk-silicon devices     

Comments, with reply, on `Performance of coherent opticalreceivers' by J.R. Barry and E.A. Lee

The commenter shows how nonzero extinction ratios affect the performance of an otherwise ideal on-off keyed (OOK) communication system. In their reply, J.R. Barry and E.A. Lee point out that by using these results, experimenters will be able to state accurately what fraction of their sensitivity deviation from the ideal quantum limit is due to the nonzero extinction ratio. The ideal ten-photon quantum limit for OOK measures the number of photons per bit required by an ideal system-including the transmitter and channel, and not just the detector-to achieve a BER of 10^-9. Barry and Lee agree with the commenter that it is not a useful benchmark for detectors alone; it is, however, a useful benchmark for the entire system     

Comments on `Pulse shape, excess bandwidth, and timing errorsensitivity in PRS systems'

Several non-minimum-bandwidth partial-response-signaling (PRS) techniques were compared in the above paper (ibid., vol.COM-35, p.475-80, April 1987). One of the criteria used for the comparison was the robustness to the timing jitter. The commenter shows that according to this criterion the duobinary signaling technique is superior to the dicode scheme at low values of the roll-off factor, in contradiction to the conclusion in the original study. In reply the author agrees that the comment is correct and that unnormalized and normalized figure-of-merit plots show different aspects of timing jitter sensitivity as a function of excess bandwidth     

Comments, with reply, on `Planetary influences on electricalengineering' by R.N. Bracewell

The commenter adds an example of the influence of astronomy on electrical engineering to those that appear in the above-titled paper (see ibid., vol.80, no.2, p.230-237, 1992). He points out that except for a factor of approximately 10, Bode's law of planetary orbit radii is quite similar to Bessel function radial units of first-order cylinder functions of the first kind, such as appear in harmonic microwave cavities used in electrical engineering. In his reply, the author presents another example of an electrical topic with roots in astronomy: the problem of determining dynamical stability by examining the differential equation of the motion, a basic part of modern stability and control theory in the electrical engineering curriculum that was first attacked by Isaac Newton in connection with the stability of planetary orbits     

Comments, with reply on `Magnetotelluric exploration forhydrocarbons' by A.S. Orange

An oversimplification of the treatment of the incidence angle and of the source dependence of magnetotelluric surface impedance in a previous treatment of magnetotelluric exploration of hydrocarbons (see ibid., vol.77, no.2, p.287-317, 1989) is pointed out. The author of the original paper agrees with the comments and recognizes the paper's lack of rigor. He points out that the intent of the paper was to present a broad overview to a general audience and to focus on the utility and applications of the method     

Comments, with reply, on `Matrix formulation for Dolph-Chebyshevbeamforming' by A. Zielinski

In the above-titled letter by A. Zielinski (see ibid., vol.74, p.1799-1800, Dec. 1986), two linear sampling intervals are proposed for the matrix determination of Dolph-Chebyshev weights, for the case of a linear array consisting of an even number of elements. However, it appears that the use of either interval results identically in a rank-deficient coefficient matrix, such that a solution is not possible. The commenter proposes nonlinear sampling intervals for both an even and odd number of elements which do not appear to display this difficulty. Zielinski replies that there is a typographical error in the original paper. This correction has been published (see ibid. vol.77, p.934, Jun 1989)     

Comments, with reply, on 'Cumulants: a powerful tool in signal processing' by G.B. Giannakis

A counterexample is presented to show that the solution to the stochastic realization problem given in the above-titled letter (ibid., vol.75, no.9, p.1333-4, Sept. 1987) does not apply, in general, to the infinite impulse response systems. The author expands on the comments.<>     

Comments, with reply, on `Synchronization and relativity' by G.M.R.Winkler

The commenter asserts that the above-titled paper by G.M.R. Winkler (see ibid., vol.79, no.7, p.1029-39, July 1991) perpetuates a fairly widespread fundamental error by using a semipopular explanation for the dissemination of time in rotating frames of reference. The misconception that one must employ general relativity to describe kinematic phenomena (synchronization, Sagnac effect, etc.) in accelerated systems, even in flat space-time, is exposed by the commenter, who proposes a way to resolve the problem. Winkler replies that he shares the commenter's concerns but does not necessarily agree with the way he proposes to resolve the issues, and he presents arguments to support his position. In a further comment, the commenter disagrees and offers further discussion     

Comments, with reply, on 'An exponential second-order probability-density function' by V.P. Pyati

The commenter discusses the claim by the author of the above-titled letter (ibid., vol.75, no.11, p.1548-9, Nov. 1987) that the joint probability density function (JPDF) used previously by the commenter is unrealistic in the limit of widely separated (time-series or rough-surface) points. The nature of the use of the JPDF in the physical optics integral for very-rough-surface scatter avoids the limit where this problem occurs, so the scatter results based on the commenter's model remain valid. He further observes that the exponential JPDF given by Pyati is neither simple nor easy to manipulate. Pyati presents some counterarguments.<>     

Comments, with reply, on `Optimal respiratory controllerstructures''

The commenter points out several defects in the modified respiratory controller model described in the above paper (ibid., vol.BME-33, p.677-80, 1986). He argues that: (1) the assumption of additive response characteristics is problematic, (2) the proposed additive controller structure is inherently redundant and unnecessary, and (3) these ambiguities aside, an alternative additive-type optimal controller model exists which has a different neuronal structure than the one described by the authors, i.e. the additive controller structure is not unique. The authors provide a counterargument     

Comments with reply on `Evaluating fault trees (AND and OR gatesonly) with repeated events' by K.S. Brown

The commenter summarizes his own approach to fault tree analysis (see ibid., vol.R-36, no.2, p.250-54, 1987) and wonders why K.S. Brown (see ibid., vol.39, no.2, p.226-35, 1990), did not compare his own approach with the commenter's approach. He asks Brown to comment on the latter and raises a number of questions about Brown's paper. Brown, in reply, points out the limitations of the commenter's approach and responds to each of the points he raises     

Comments, with reply, on `Fast interpolation algorithm using fastHartley transform' by J.I. Agbinya

The commenters point out and correct several errors in the original letter (ibid., vol.75, no.4, p.523-4, Apr. 1987). In replying the author states that some suggested corrections to the gain of the low-pass filter and the kernel of the cas function, respectively, are in order. But the author points out that Hsa et al. have introduced a further error into the cas function themselves. The correct version is shown     

Comments, with reply, on `Ada technology: current status and costimpact' by P. Bhansali

The commenter argues that, in the above-titled paper by P.V. Bhansali (see ibid., vol.79, p.22-29, 1991): (a) data reported to be supported by a footnoted reference are not in the referenced article; (b) data in a referenced article that are at variance with the conclusions reached in the paper were not included; and (c) when these data are included, the conclusions reached concerning Ada productivity are not supported by the facts. Bhansali replies that the main objective of conducting the cost/benefit analysis was to come up with a parametric model so that individual organizations can utilize their data and determine the effectiveness of using Ada. The most difficult part of conducting the cost analysis was obtaining consistent data. Because of data limitations the conclusions was equivocal: Ada could represent a potential savings over the entire life-cycle     

Comments, with reply, on 'Solving the generalized eigenvalue problem with singular forms by M.D. Zoltowski

The above-titled letter (ibid., vol.75, no.11, p.1546-8, Nov. 1987) proposed two numerical algorithms for the solution of generalized eigenvalue problem Ax= lambda , where it is assumed that the matrix B is Hermitian. It was claimed by the author that standard algorithms are not available if the matrix B is singular. The commenters draw attention to the literature in this field based on the QZ algorithm. The author thanks the commenters for supplying the information but points out to prospective readers certain caveats with regard to use of the QZ algorithm.<>     

Comments, with reply, on `On the hazard rate of the lognormaldistribution' by A.L. Sweet

In the above-named work (see ibid., vol.39, p.325-8, Aug 1990), A.L. Sweet calls attention to some incomplete or inaccurate representation of the hazard rate of lognormal distributions. His principal concern apparently is that these representations may lead modelers to overlook an important property of the lognormal distribution: its hazard rate increases and then decreases as a function of elapsed time. He also provides approximations for the location and value of the hazard rate maximum. The commenter observes that, as theoreticians are likely to have ready access to expanded tables of the normal distribution and unlikely to have need for the approximations, the Sweet paper was aimed at practitioners dealing with actual data. The commenter cautions such practitioners that some serious problems, unrelated to table availability or approximation accuracy, arise if σ_1n(T) is high or low. In his reply, Sweet remarks that the commenter is correct in assuming that Sweet's work was aimed at practitioners-specifically, at those being introduced to the lognormal distribution for the first time by reading about it in the books referenced in Sweet's article     

Comments, with reply, on 'A real-time high-resolution technique for angle-of-arrival estimation' by J.P. Reilly

See ibid. vol.75, no.12, p.1692-4 (Dec. 1987). The commenters explain why they disagree with some assertions in the above-titled letter and point out some omissions that they consider important. In particular, they maintain the systolic array of W.M. Gentleman and H.T. Kung cannot be used for subspace estimation in high-resolution methods. The author points out some alternative interpretations that present the original work in a different light.<>