Index of content:
Volume 23, Issue 5, 01 May 1952
23(1952); http://dx.doi.org/10.1063/1.1702240View Description Hide Description
A sensitive and accurate method for determining the real part of the dielectric constant of gases at microwave frequencies is described. The method eliminates some possible sources of systematic error inherent in earlier methods.
The dielectric constant of dry air at N.T.P. has been determined as 1.000569, at a frequency of 3000 Mc/sec. This is in good agreement with other results, at 24,000 Mc/sec, at 9000 Mc/sec, and at low frequencies, thus verifying that the dielectric constant is independent of frequency up to nearly 30,000 Mc/sec An earlier result suggesting the contrary was found to be in error because of insufficient drying of the air.
23(1952); http://dx.doi.org/10.1063/1.1702241View Description Hide Description
We consider the shorted‐line method as applied to general samples, not restricted to quarter‐wave, half‐wave, thin, or low loss. The complex dielectric constant can be obtained by combination of two measurements, the sample being followed by short and open circuit, respectively. The equation is simple, nontranscendental, and unambiguous (Eqs. (5), (11), (19), (22)). It lends itself to complete graphical representation (Table I, Fig. 2). When only a single measurement is made, as in conventional practice, a change of variable leads to great simplification of detail ((14), (15), (16), Figs. 2, 3, 4). With due regard to the nature and magnitude of experimental errors, we analyze the conditions for optimum graphical representation, and give a reasoned determination of scale ranges. Large scale curves have been constructed, but only reductions are given here.
23(1952); http://dx.doi.org/10.1063/1.1702242View Description Hide Description
1. The high limiting contact voltage of 40 to 50 volts in contacts between siliconcarbide members (single crystals of the black‐green variety) is explained on the basis of the known temperature‐voltage relation.
2. Practically clean siliconcarbide possesses a superficial layer which functions as a potential hill for the electrons. Because of its thinness the hill is penetrable for the electric current by means of the tunnel effect. The tunnel resistance per unit area was investigated with terminal areas of the order of 2 mm2. Being constant at small voltages, it drops down through some orders of magnitude within the voltage range of 0.5 to 5 volts, showing that the work function for emission of electrons into the barrier material from the adjacent metal is of the order of 3 electron volts. From the amount of the tunnel resistance it is concluded that the thickness of the potential hill is of the order of 10 to 20 Angstrom units. The simultaneity of applied voltage and tunnel current is demonstrated.
23(1952); http://dx.doi.org/10.1063/1.1702243View Description Hide Description
The hypothesis is suggested that initiation of high voltage breakdown in vacuum is due to traversal of the high voltage gap by a clump of loosely adhering material. The implication of this hypothesis for uniform‐field gaps is that the breakdown voltage is proportional to the square root of the gap length. A summary the literature is presented which supports this conclusion for a range of voltages from 20 kilovolts to 7 megavolts, and for a range of gap distance from 0.2 mm to 6 meters. Additional qualitative evidence is presented which tends to support the proposed hypothesis.
23(1952); http://dx.doi.org/10.1063/1.1702244View Description Hide Description
A method of estimating magnetization losses due to eddy‐currents in solid cylindrical cores of ferromagnetic material is given. The conception of complex permeability is introduced to show the effects which hysteresis has upon the degree of flux penetration. A formula is given for the effective permeability which should be used in the classical treatment in which nonlinearities have been ignored.
An arbitrary case in which the permeability changes across the core section is considered and the solution is extended to a core which is homogeneous, but whose permeability changes with the degree of magnetization. In this way errors due to a nonlinear B‐H curve are reduced and the complex permeability notation is shown to lead to a more reasonable Steinmetz relationship.
The eddy‐current anomaly is mentioned and it is noted that the theory developed is suited to a consideration of the anomaly as explained in terms of time‐lag effects.
23(1952); http://dx.doi.org/10.1063/1.1702245View Description Hide Description
A possible source of confusion between the concept of information content and entropy in the theory of information is discussed and resolved. Information theory is then applied to the problem of taxonomy or classification of data and several models are discussed, representative of various possible methods of filing data with the purpose of determining the optimum size of filing‐unit in relation to the given data. The possible fields of application envisaged include the classification of faults in a system, the statistical analysis of data or the indexing of books, letters, etc.
23(1952); http://dx.doi.org/10.1063/1.1702246View Description Hide Description
An article by Birkhoff, MacDougall, Pugh, and Taylor (see reference 1) presented hydrodynamic theories of jet formation and target penetration by explosives with lined conical cavities. However, it was unable to explain satisfactorily why the jets produced are several times as long and, therefore, several times as effective as the steady‐state theory predicts. It is shown here that these difficulties are overcome by assuming a variable instead of a constant collapse velocity for the walls of the conical liner. The variability in the collapse velocity produces a surprisingly large change in the process of jet formation.
23(1952); http://dx.doi.org/10.1063/1.1702247View Description Hide Description
Experimental evidence is presented that verifies the assumptions and conclusions of the article, ``Theory of Jet Formation by Charges with Lined Conical Cavities,'' by Pugh, Eichelberger, and Rostoker. A graphical method of predicting the shape of the jet and slug at every instant is presented.
23(1952); http://dx.doi.org/10.1063/1.1702248View Description Hide Description
23(1952); http://dx.doi.org/10.1063/1.1702249View Description Hide Description
The subject is an unconventional method of heating and melting metals without a crucible, by suspension in space with an electromagnetic field. Operating conditions for certain cases are given. The results obtained by means of the new technique encourage the thought of melting, purifying, alloying, and agitating of inert and reactive metals without resort to crucibles, and thereby avoiding the contamination of reactive metals by crucible materials. Preliminary results with various forms and masses of metal are described. Considerations concerning the atmosphere in which levitation occurs are included.
23(1952); http://dx.doi.org/10.1063/1.1702250View Description Hide Description
A description is given of a system made up of experimental electrodes and an oscilloscope by means of which the potential across the electrodes can be recorded with a time resolution of about 10−9 sec and a potential sensitivity of 1‐trace width per volt. The closure of the electrodes to produce a short arc is synchronized with the oscilloscope sweep so that the beginning of the arc is photographed.
As an arc starts the potential across the electrodes decreases more or less gradually from the applied voltage to a steady value characteristic of the metal of the electrodes. The course of this change is extremely variable as is also the time over which the change is spread. The average value of the time appears to vary with circuit inductance and with the nature of the electrode surfaces. For inactive silver surfaces and an inductance of 0.10 μh the average value of the time is about 0.007 μsec, and for active surfaces and the same inductance 0.011 μsec. For active surfaces and an inductance of 5 μh the average value of the time is 0.02 μsec.
The electrode separation at which an arc strikes is determined from the oscilloscope traces and from a correction for the height of the mound of metal thrown up by the arc. For active silverelectrodes the average separation (at 40 or 45 volts) corresponds to a gross electric field of 0.8×106 volts/cm, and for inactive silverelectrodes to a field of 2.3×106 volts/cm. These are probably better values than earlier measurements of these fields. There has not yet been any success in interpreting these phenomena in terms of fundamental processes.
23(1952); http://dx.doi.org/10.1063/1.1702251View Description Hide Description
A magnetically focused space‐charge cloud of the type found in the smooth‐anode magnetron is examined to determine whether a perturbation of the equilibrium condition will grow. A field analysis is carried out in which radial admittances are matched at the edge of the cloud. The solutions for the characteristic frequencies are complex, indicating that the disturbance grows with time. These instabilities occur at multiples of the Larmor frequency.
The solutions indicate that electrons oscillate with increasing swings, a process which absorbs energy. Instability therefore gives rise to some anode current even when the tube is ``cut off'' though rf power is not made available by the process.
23(1952); http://dx.doi.org/10.1063/1.1702252View Description Hide Description
A method of solving electron beam problems is described which takes into account the thermal velocity spread. The method is based on the Liouville theorem. Two such problems are solved by means of a power series. In the case of small‐signal velocity modulation of a drifting stream, the velocity spread distorts the shape of the standing wave of current which is produced, in addition to lengthening the plasma wavelength and the electronic wavelength. For a drifting stream with full shot noise in each velocity class at the input, a standing wave of noisecurrent is produced with deep minima. The size of the first minimum is found to increase as the square of frequency and with increasing velocity spread. That full uncorrelated shot noise can produce a standing wave of convection current is somewhat surprising. The present analysis demonstrates that space charge is the important factor in determining the behavior of noisecurrents in electron streams even at high frequencies, and that velocity spread plays a less important role.
23(1952); http://dx.doi.org/10.1063/1.1702253View Description Hide Description
Measurements of magnetically induced velocity changes in polycrystallinenickel rods are discussed. Shear and compressional ultrasonic propagation modes are used throughout the frequency range of 1–10 megacycles. The polycrystallinemeasurements are compared to recently published single nickel crystal values. The ΔEeffect is found to decrease with increasing frequency throughout the megacycle region, showing general agreement with low frequency measurements.
23(1952); http://dx.doi.org/10.1063/1.1702254View Description Hide Description
With a view to studying the mechanism of heat conduction in low conductivity alloys, a method has been devised by which the thermal conductivity of relatively small samples (⅛‐ to ¼‐inch diameter, 1 to 2 inches long) of various materials can be measured in the temperature regions obtainable with liquid nitrogen, liquid hydrogen, and liquid helium.
Preliminary measurements on several commercial alloys (monel, inconel, and stainless steel) gave Wiedemann‐Franz ratios several times greater than the theoretical value of 2.45×10−8 watt‐ohm/deg2, the deviation being greater for annealed than for cold‐worked specimens. This has been interpreted in terms of appreciable lattice conduction of heat in these alloys.
Following these results, samples of an alloy of 90 percent copper 10 percent nickel were prepared with varying amounts of cold‐work and with different grain sizes. Results with these samples were similar to those obtained with monel and inconel and confirm the hypothesis of lattice conduction. They also give a qualitative indication of the effectiveness of cold‐work in limiting the lattice conduction of heat.
23(1952); http://dx.doi.org/10.1063/1.1702255View Description Hide Description
In a previous paper [The Input Impedance of an Antenna Over a Finite Ground Screen, Technical Report No. 119 (Cruft Laboratory, Harvard University, November 30, 1950)] the author treated the problem of the input impedance of an antenna erected over a large circular screen. Using the formulation developed therein, an expression is obtained for the entire radiation pattern of an antenna over a large circular screen.
The most important difference between the pattern of an antenna over a large screen and an antenna over an infinite screen is the presence of lobes over the antenna. From the expression for the radiation pattern, simple formulas are obtained for: (a) the number of lobes, (b) the angular position of the lobes, and (c) the angle, βmax, within which the lobes occur. One important result obtained is that
23(1952); http://dx.doi.org/10.1063/1.1702256View Description Hide Description
Since argon is prepared by fractional distillation from liquid air, the greatest impurity found in it is nitrogen whose boiling point lies close to that of argon. For this reason it is believed that some of the previous measurements on supposedly pure argon were actually made on nitrogen contaminated argon. By adding controlled amounts of nitrogen to pure argon, the drift velocity curves of various authors can be duplicated.
23(1952); http://dx.doi.org/10.1063/1.1702257View Description Hide Description
23(1952); http://dx.doi.org/10.1063/1.1702259View Description Hide Description