Volume 12, Issue 9, 01 September 1941
Index of content:
12(1941); http://dx.doi.org/10.1063/1.1712958View Description Hide Description
This article reports an investigation of the current and voltage characteristics of ``ignitor'' electrodes in a mercury pool rectifier (ignitron). Experiments were made on square rods of smooth homogeneous materials of various sizes and resistivities with the following results: (1) A frequency distribution curve is found for the voltage required by the ignitor which is approximately a Normal or Gaussian distribution. (2) The total current required for the ignition of an arc, divided by the perimeter is a constant, for a given material. (3) The total current required per unit length of perimeter times the square root of the resistivity is a constant. (4) For identical ignitors operated in parallel, the total current required is equal to the square root of the number of ignitors times the current required for one ignitor. (5) An expression is derived for the fraction of the total current which is useful in starting the arc and this is found to be inversely proportional to the square root of the width. This derivation is verified by the behavior of ignitors operated in parallel.
12(1941); http://dx.doi.org/10.1063/1.1712959View Description Hide Description
By coordinating our information on viscosity,diffusion,melting, and other rate and thermodynamic properties, we arrive at a detailed picture of liquid structure. Thus we find that a liquid is best thought of as a solid to which a large number of empty equilibrium positions are added. In fact the expansion on melting, as well as the expansion with a rise in temperature, arises almost entirely from this introduction of new equilibrium positions, and only to a minor extent from lattice expansion. We shall obtain information as to the number, size and energy of formation of these empty lattice points.
12(1941); http://dx.doi.org/10.1063/1.1712960View Description Hide Description
The purpose of the paper is to indicate a method by means of which elastic constants (bulk and Young's moduli) and viscosity can be computed, one from the other, under conditions in which the direct determination of one of the data required is difficult or not feasible, and the other of the two quantities in question is known.
12(1941); http://dx.doi.org/10.1063/1.1712961View Description Hide Description
Starting with the set of linear equations that determine the complex currents of a general n‐mesh circuit, a method is presented for the numerical solution of the equations to determine the various mesh currents of the system.
The method presented has computational advantages over the conventional methods usually used in that only real numbers are involved in the various numerical operations and only the evaluation of second‐order determinants is involved. The operations required are those of matrix multiplication and are most conveniently performed with a calculating machine. The paper is concluded by the solution of a general Y‐Y circuit with a neutral return as an example of the method presented.
12(1941); http://dx.doi.org/10.1063/1.1712963View Description Hide Description
12(1941); http://dx.doi.org/10.1063/1.1712964View Description Hide Description
This is a continuation of the work reported in 1934. The greater part of the paper is devoted to a discussion of the properties of iron‐nickel‐cobalt alloys, especially the composition Fe 54 percent, Ni 31 percent, Co 15 percent—called fernico—including sensitiveness to impurities and polymorphism. Three new glass‐metal combinations are described, the metals being 42 percent nickel‐iron, 26 percent chromium‐iron, and pure iron, respectively. The strength of the glass‐metal bond is discussed briefly.
12(1941); http://dx.doi.org/10.1063/1.1712965View Description Hide Description
An examination of the interrelation between the sine series solution (for a slab with sealed periphery) and the error function solution (for a semi‐infinite body) of the fundamental Fourier equation has yielded the following results: (1) Distribution functions according to the two solutions begin to diverge soon after diffusion has begun. (2) The error function yields an excellent simple approximation, valid until half the material initially present has been lost, for the loss of material from a slab with sealed periphery. (3) A simple, but extremely good, approximation is given for an infinite exponential series in the region of slow convergence. (4) Two simplifications in the calculation of diffusion constants have been given. (5) While the discussion has been restricted to diffusion it is applicable, with obvious modifications, to other cases of flow governed by the fundamental Fourier equation.