Abstract :
Review of the Subject. — There are many engineering problems which can be solved step by step, and there are others which contain two or more inter-dependent unknown quantities and therefore cannot be solved step by step. For example, the diameter of a short transmission shaft, mainly subjected to torsion, can be computed directly for a desired shearing stress; then the pulley can be designed, the belt for the pulley, etc. This is a step-by-step solution. On the other hand, let it be required to determine the diameter D and the length L of a cylindrical tank, to hold a given quantity of liquid and to possess a given cooling surface. — This problem either requires-two simultaneous equations for D and L, or else has to be solved by trials. The problem of determining theoretically the internal voltage drop and voltage regulation of a given synchronous generator at a given load is one of the second kind, that is, several conditions must be satisfied simultaneously. The problem of predetermining the armature current and the power factor of a given synchronous motor at a given excitation and load is of a similar nature. The purpose of the kinematic device described in the paper is to enable one to solve these two problems on synchronous machines mechanically, almost without any mathematics and without constructing vector diagrams. The device can be used even by a person who does not know its theory, just as thousands of computers who use the slide rule could not explain the theory of its scales. A knowledge of the performance characteristics of a synchronous machine is of considerable practical importance, both to its designer and to the user; it has therefore been deemed advisable to simplify and to standardize the procedure so as to make the process as nearly automatic as possible. This should give more time to progressive engineers to think about other problems connected with synchronous machinery, problems which are only partly solved, such as the tem- erature rise, insulation, transient phenomena, etc. Two factorsin particular complicate the performance of synchronous machines (especially with salient poles), namely the armature reaction and the variable saturation of the magnetic circuit. The effect of the armature reaction is to weaken (or to strengthen) the field flux and also to shift it sidewise. The method of predetermination of voltage regulation given in the A. I. E. E. Standards is admittedly an approximate one, and in designing the Blondelion a more accurate method of taking these two components of the armature reaction into account has been used. This method is known as Blondel´s theory of two armature reactions. The curves in Figs. 10 to 13 show a close check between the experimental curves of certain machines and the points read off on the kinematic device. The Blondelion can be readily simplified for use in accordance with the A. I. E. E. method. The variable saturation of the magnetic circuit is taken care of in the device automatically, by means of a proper linkage. To illustrate the principle upon which the Blondelion is built, let us take again the above-mentioned tank problem and see how a mechanical device could be made for its solution. Let various values of tank diameter D be marked on a certain scale and let the values of tank length L be marked on another scale. Assume that each scale has an index which can slide along it and that these two indexes are kinematically so interconnected that a pointer indicates directly the corresponding volume of the tank on a third scale. The arrangement is somewhat indefinite in that the same volume can be obtained with an infinite number of combinations of values of D and L. Let now the same indexes be also interconnected by another kinematic linkage, such that another pointer, on a fourth scale, gives directly the area of the cylinder. Here again, an infinite number of combinations of D and L will give the same area. In order, however, to make the thir
