[Illustration: Fig. 704.--Method of locating short circuits between adjacent armature coils. Fasten a monkey wrench to the rim of the pulley, or a crank to the shaft. Now, excite the fields, and, to make the effects more marked, connect the coils in parallel. When this has been done it will require considerable force to rotate the armature, and then it will move quite slowly, except at one position. When this position has been found, mark the armature at points in the center of the pole pieces at points A and B and at both ends of the armature. The explanation is that both halves of the armature oppose one another at this position; but when not at these points a continuous circuit is formed, and the resultant magnetic effect is considerable. The "cross" or "short" circuit is nearly always found on the commutator end in the last half of the winding, where the wires pass down through the first half terminals. This applies to an unequal winding. In armatures where the windings are equal, it is as liable to occur at one point as at another. With this method a defect can be found and remedied in a few moments, for it has always been a simple matter to repair it when discovered. These results can be observed in a perfect armature by connecting the opposite sections of the commutator.] Short Circuits between Sections through Frame or Core of Armature.--Detection of this fault can be effected by the methods described above, and by disconnecting the whole of the armature coils from the commutator and from each other, and testing each separately with a battery and galvanometer coupled up as in fig. 705, one wire being connected to the shaft and the other to the end of the coil under test. As a rule, there is no way of remedying this fault other than unwinding the defective coils, reinsulating the core, and rewinding new coils. [Illustration: Fig. 705.--Method of locating short circuits between coils through armature core. The galvanometer, battery and coil to be tested are connected in series as shown, and then the unconnected terminal of the galvanometer is brought into contact with the shaft. If then some portion of the insulation of the wire has been abraded or destroyed, thus bringing the bare wire into contact with the metal core, as at A in the figure, the needle of the galvanometer will be deflected since a closed circuit is formed through the core and wire. If the insulation be perfect, the needle will not be deflected. It will thus be seen that in the conductivity test (fig. 700) it is necessary that the needle should be deflected, or turned, to prove that all is right, while in the insulation test the converse holds good; if the needle be deflected, it proves that the insulation is broken down.]