Salient-Pole Machines

The first electromechanical rotating machines had salient poles and did not make good motors

It was not long after it was shown that an electrical current produced a magnetic field that could move magnets and seemed to circle around a wire carrying current that experimenters searched for evidence of movement produced by electricity. Michael Faraday made a wire revolve around a magnet, and a magnet revolve around a wire in September 1821. In the same year, Peter Barlow caused a sold toothed disc between the poles of a magnet to revolved continuously, the moving contact with the teeth provided by mercury. Faraday was content with his homopolar disc generator and did not investigate electrical machines more deeply.

In 1830, dal Negro contrived an oscillating magnet that produced rotary motion through a pawl and ratchet. Joseph Henry, using wires dipping into mercury to make contacts, also made an oscillating device in 1831. Pixii rotated coils at the end of a permanent magnet to generate alternating currents, and at Ampére's suggestion seems to have invented the commutator to rectify the output, as an improvement on an earlier cam-operated switch. Pixii's machine could easily be inverted to make a motor, and in this form the small salient-pole machine became quite popular. Other early motors were made by Christie and Ritchie, that were very much like the type described below, with a permanent magnet field and wound rotating armature. These motors only rotated, and did no useful work. Theodore Sturgeon (inventor of the electromagnet) constructed a 4-pole machine that turned a roasting spit through a worm and spur gear in 1832. In Holland, Stratingh made two-pole machines very much like the one illustrated below, and used them to drive small cars and boats in 1835.

In America, Thomas Davenport developed a 4-pole motor, and received both U.S. and British patents in 1837. He made a small electric train that ran on a circular track, a drill press, and, most ambitiously, began work on an electrically driven printing press. His business associates ran off with the money, and further work under great difficulties led to utter failure. At the beginning of the 1840's, Robert Davidson in Scotland experimented with electric boats and locomotives, and actually tested a battery-powered locomotive. His electromagnets simply attracted iron bars to produce motion. These activities attracted public attention on both sides of the Atlantic, but the practical results were negligible. Jacobi, in St. Petersburg, constructed electric boats and operated them on the Neva. In the United States, Moses Farmer, Thomas Hall, Charles G. Page and others all made experiments and gave public demonstrations through the 1840's and 1850's, again with no practical outcome.

Small motors of the type found in the early experiments are still sold for play and school demonstrations, purporting to teach the principle of the electric motor and generator. These devices do indeed have a lot to teach, but they do not really show how the electrical machines we now have function. In this paper, I will try to explain how these small motors actually work, and why they did not lead to practical motors. In fact, it was not until about 1875 that modern, efficient motors were eventually developed, after the fundamental principles became better understood. However, this was only about forty years--we have been looking for practical fusion power for over fifty now.

One part of a salient-pole motor was a magnet, a permanent magnet in most of the early machines, or else an electromagnet, and usually in a U-shape, with two poles, N and S, that were indeed salient, or "sticking out." This part was later called the field, since it supplied the essential magnetic field. A piece of iron was arranged to rotate between the poles of the field. Permanent magnets had begun as chunks of magnetite with iron pole pieces called the armature, and the same name was used for the "keeper" of a permanent magnet. Therefore, this rotating piece of iron was also named the armature, from its position near the poles of the field. The armature would naturally take up a position between the poles that would result in the greatest magnetic flux, a position of minimum reluctance, and would be held there, restrained from rotation.

The armature was wound with many turns of wire, so that a current through this wire would make the armature an electromagnet. The idea was that the current should be in such a direction when the armature was approaching the poles that the magnetic poles of the armature and field should attract. When the armature was receding, the magnetic poles should repel. This, of course, required a reversal of the current direction about the time when the armature was in the position of minimum reluctance. The critical invention was the commutator, that would automatically commute, or change, the connections to the armature so that the field would reverse. This consisted of segments that rotated with the armature, to which the external circuit was connected with sliding brushes. These actually were brushes of phosphor bronze wire; carbon brushes did not come until much later. This also explains why blocks of graphite are now called brushes.

These things are shown in the sketch at the right. The field, armature, commutator and brushes are pointed out. Similar parts have appeared on all DC machines since then. The field is shown as a permanent magnet, although it could just as easily be an electromagnet supplied from the external power source. Connections to the brushes and commutator are shown, and can be used to find out the directions of the magnetic fields created by the currents. Current flows from + to -, and if the thumb points in the direction of the current, the magnetic field curls around like the fingers. At the left, there is no current, because the brushes are over the insulating segments of the commutator, and the armature is held in the position of minimum reluctance.

Suppose now that the armature is rotated a little in the direction of rotation of the motor (clockwise, with the connections shown). The brushes make contact, and a current flows that reverses the direction of the magnetic flux in the armature, and makes its poles the same polarity as those of the field. If we had rotated the armature the other way, the current would be reversed, and the armature poles would be of opposite polarity to the field poles. In either case, the strong magnetic force impels motion in the same direction, clockwise. The armature is given a strong double kick as it moves past the poles and the armature windings are commutated. It is kicked at both ends, as well. From its inertia, it rotates half a turn until the same thing happens. There are four kicks for every rotation, and the motor rotates very well indeed. It is a good exercise to work out the directions of the currents, magnetic fields and forces for various positions of the armature. The forces really should be found from an accurate map of the magnetic field as the armature rotates, but it is easier to use the concept of force between magnetic poles, as was done at the time.

This motor runs so nicely that one thought immediately of scaling it up so it could do useful work. The result was total failure, for two principal reasons. One is the short range of magnetic forces. A small machine got four kicks that extended over reasonable fractions of the rotation. A large machine also got four kicks, but over a very limited distance, so that the actual work done, force times distance, did not scale up with the machine. One help for this was to multiply the number of poles, and this was very commonly tried. The second reason, and this is the more important one, was that for most of the time the external power source was merely forcing large currents through a small resistance without doing anything useful. This great inefficiency not only wasted expensive zinc in the batteries, but also heated the motor excessively.

To appreciate this latter effect, remember that a changing magnetic field induces an electric field in the direction that would cause any currents flowing as a result to oppose the change in magnetic field. When the armature is being strongly attracted or repelled by the field, the magnetic flux is being changed in the armature, and a voltage is induced in the armature windings, just as in a transformer. This voltage opposes the applied external voltage, and is called a back emf (emf = electro-motive force). The external batteries then work against this back emf, and the work they do on it is converted to mechanical work on the rotating armature. This conversion of electrical energy to mechanical energy is 100% efficient. This is easy to see in modern machines, but was impossible for the early workers to appreciate, since even the concept of energy had yet to be developed.

Therefore, the salient-pole motor had only four small peaks of back-emf per rotation, and only the work done by the external power source on these peaks was effective. All the rest of the current was wasted, the energy converted to heat. Incidentally, conversion of electrical energy to heat is also 100% efficient. Even today, however, small motors can also afford to waste energy if the heat is not excessive (small motors cool more easily than large ones) and power cheap enough. Large motors, however, must be designed so that the external power source always works against a steady back-emf so that the efficiency is close to 100%, for the sake both of economy and heat. The small motor we described does not start itself, but this can be overcome, so that small salient-pole motors can have large starting torque. One way to do this is to wind the armature with an odd number of poles so there is no point of minimum reluctance, and to excite the windings to give the maximum torque first on one side, then on the other, of the armature.

Many investigators tried to improve the efficiency of the salient-pole motor, among them James Prescott Joule of the conservation of energy and Jacobi, an eminent mathematical physicist, only concluding that such motors were so inefficient as to be impractical, if only because they wasted expensive zinc. In fact, these investigations led Joule and Jacobi to the study of the conservation of energy. Even when cheaper power from dynamos was available, such motors did not become practical, because they had the fatal flaws we have noted.

Most later efforts concentrated on the mechanical problems rather than the electrical ones, usually trying to find some way to extend the very short range of strong magnetic forces. The result was a collection of exotic machines with eccentric motions and various forms of leverage, none of which did any good.

As a converter of mechanical into electrical energy, the salient-pole machine was much more successful. If the armature is simply rotated in the field, the alternating magnetic flux in the armature core generates a pulsating DC when rectified by the commutator. If the connections to the armature are made by brushes sliding on continuous slip rings, then the output is AC. These outputs have peaked waveforms, not the steady DC nor sinusoidal AC that are ideals, but the DC machine can serve well for electroplating, and the AC machine for arc lights. Since the AC machine does not require a commutator, the field can be made the rotating element and the armature the stationary one, making insulation and connections much easier. Dynamos and alternators of this kind were relatively successful long before good motors were developed. The peaked waveform of the output meant that there was considerable heating, so these machines ran hot, but at least they ran. Some were used only to give shocks as a parlor trick, but others carried out the more respectable duties of lighthouse illumination and electroplating.

Another device for converting electrical to mechanical energy is the solenoid, but the limited and unidirectional nature of the force is a great inconvenience. As the iron core of the solenoid moves, the magnetic field changes, and this causes a back emf as in a motor. Again, the only useful work by the current sources is that done against the back emf, and the sources expend most of their energy making heat. Although solenoids have important applications, the efficient conversion of electrical to mechanical energy is not one of them. Reciprocating solenoid motors were, however, tried. They were unsuccessful, as can be imagined.

The changing currents produce another effect, self-inductance, in the armature windings. If these windings are made of many turns of fine wire, with the aim of making a higher resistance that will not create so much heat (heating is proportional to the square of the current), the changes in the magnetic field are delayed, and voltages are induced in the windings that can cause sparking at the commutator and breakdown of insulation. The self-induction voltages are quite different from the back emf, and do not represent energy transformation (except to and from the magnetic field). With all of the effects we have mentioned, it is no wonder that inventors were baffled in their efforts to produce an efficient motor.

At about the same time in the 1870's, Gramme and Siemens discovered the secret of efficient electrical machines. They were actually making dynamos, but soon observed that their dynamos also made excellent motors. A Gramme machine is shown in the diagram at the left. The field and the armature now form a magnetic circuit of low reluctance that remains constant as the armature turns. The Gramme armature was a ring or hollow cylinder of iron with the continuous winding encircling it. The wires ran lengthwise on the outside of the ring, returning on the inside. The wires on the outside were actually in the magnetic field, which they "cut" when rotated as a dynamo to generate the emf according to E = Blv, and on which forces were exerted when they carried a currrent (whether as a dynamo or a motor) according to F = BlI. In this case, it is easy to see the electromagnetic interactions, and how they conserve energy. Pacinotti was the first to make a ring armature, around 1864, but he did not develop the idea, and it was left to Gramme to elaborate the concept.

The armature turns are connected to the segments of the armature, with one or more turns connected to each segment. The brushes are placed so that the coils short-circuited by them are out of the magnetic field and so are experiencing no induced voltages. This eliminates sparking at the commutator, with all the troubles that it brings. The emf increases equally on both sides of the armature, and the brushes are at the points where it is a maximum. Although the armature winding is a closed loop, no current flows around it because the voltages balance. As the armature rotates, the only thing that happens is that wires move in the constant flux produced by the field.

Siemens's contribution was to rearrange the armature conductors so that a conductor returned not inside the ring, but on the surface under the opposite pole, where the force or induced voltage would be the same. This halved the resistance of the armature and made it much easier to wind, since pre-wound coils could merely be laid on the surface, not threaded in and out. The armature conductors could also be connected in many advantageous ways. For these reasons, the Siemens drum armature is now used universally. The Gramme armature is very useful for explanations, however, since the armature connections are clear. In the diagram, assume your own polarity for the field and work out which armature terminal will be positive when the armature rotates in one direction or the other, or which way the armature will turn when supplied by current. Note that to reverse the motor, either the armature or the field connections must be reversed, not both.

Both Siemens and Gramme initially committed the error of making the armature resistance too high, because of a misinterpretation of the conditions for maximum power transfer. See Jacobi's Theorem for an explanation. This led to efficiencies not much above 40%. Edison, around 1880, showed that low armature resistances made efficient machines.

Disappearance of salient poles in the new machines made a new explanation of dynamo and motor action necessary, in terms of conductors in magnetic fields instead of attraction and repulsion of poles. Both of these analogies are partial and incomplete, but nevertheless guided and misguided experimenters. A more rigorous analogy is furnished by magnetic field plots, in which "lines of force" exhibit longitudinal tension and lateral repulsion. There are no actual lines of force, of course, but it is a useful means of expression. Very soon the armature conductors were removed from the surfaces, and placed in slots in the iron. The simple analogies failed, but the magnetic field description remained valid. Slots anchor the conductors much better, and permit a shorter air gap between the armature and the field.

When current flows in a dynamo or motor, the armature windings create a field that is different in direction from the static field. This twists the magnetic field so that the force it exerts on the armature can easily be appreciated. The lines of force in the air gap are no longer radial, but inclined in the direction of the force. This happens whether the conductors are on the surface, or are buried in slots. Another effect of the twisted field is that the point of zero field is no longer halfway between the field poles, but is shifted one way or the other, depending on whether we are talking about a dynamo or a motor. This means the brushes must be moved to eliminate sparking.

There is considerably more to a complete explanation of electrical machines, but what has been said here lays the foundation of a good understanding, both of direct current and alternating current machines. Induction motors have smooth rotors and stators; synchronous machines have salient-pole rotors and often stators as well; alternators have salient-pole rotating fields and smooth stators. The main idea is always to have the power source work against a back emf, and not dissipate its energy in heat.


  1. B. Bowers, A History of Electric Light & Power (London: Peter Peregrinus, 1982)
  2. P. Dunsheath, A History of Electrical Engineering (London: Faber and Faber, 1962). This is a good history, but oddly has little on motors.

Return to Tech Index

Composed by J. B. Calvert
Created 14 February 2001
Last revised