COMPENSATION WINDING BASIC INFORMATION AND TUTORIALS

Although the commutating pole is a good solution for commutation, it does not prevent distortion of the main-pole flux by armature reaction. The flux set up across the main-pole face by the armature mmf is shown in Fig. 8-30a.

If the pole face is provided with another winding, as shown in Fig. 8-30b, and connected in series with the load, it can set up an mmf equal and opposite to that of the armature. This would tend to prevent distortion of the air-gap field by armature reaction.

Such windings are called compensating windings and are usually provided on medium-sized and large dc machines to obtain the best possible characteristics. They are also often needed to make machines less susceptible to flashovers.

The use of compensating windings reduces the number of turns required on the commutating pole fields, and this materially reduces the leakage fluxes of the field and, in turn, the pole saturations at high currents. The ampere-turns on the commutating field are reduced by about 50% with the use of a compensating field.

This new winding may be considered to be some of the turns taken off the commutating-pole winding and relocated in slots in the main-pole faces. The number and location of the compensating slots must be carefully chosen to match, as closely as possible, the rotor ampere-turns per inch.

However, the slot spacing must not correspond closely to that of the rotor. This would cause a major change in reluctance to the main-pole useful flux every time the rotor moved from a position where the rotor and stator slots all coincided to where the rotor slots coincided with the stator teeth.

This would occur once for every slot-pitch movement. The resulting rapid changes in useful flux would cause ripples in the output voltage and also serious magnetic noise. If too few slots are used, local flux distortions occur and the compensating winding loses some of its effectiveness.

Compensation of armature reaction effectively reduces the armature circuit inductance. This makes the machine less susceptible to the bad effects of L(di/dt) voltages caused by very fast load current changes.

During manufacture, it is possible to locate the compensating winding nonsymmetrically about the centerline of the main pole. This causes a direct-axis flux, which will give a series field effect. For generator cumulative compounding, the slots must be shifted in the direction of the machine rotation.

This shift gives a motor differential compounding. The effect cannot be adjusted after manufacture. It seldom exceeds 1/2 in, and this does not materially reduce the effectiveness of the compensation.

TWO CIRCUIT OR WAVE WINDINGS OF DC GENERATORS BASIC INFORMATION

Figure 8-16 shows a wave type of coil. Figure 8-17 gives a 6-pole wave winding. Study reveals that it has only two parallel paths between the positive and negative terminals. Thus, only two sets of brushes are needed

Each brush shorts p/2 coils in series. Because points a, b, and c are at the same potential (and, also, points d, e, and f ), brushes can be placed at each of these points to allow a commutator one-third as long.

The winding must progress or retrogress by one commutator bar each time it passes around the armature for it to be singly reentrant. Thus, the number of bars must equal (kp/2) # 1, where k is a whole number and p is the number of poles. The winding needs no equalizers because all conductors pass under all poles.

Although most wave windings are 2-circuit, they can be multicircuit, as 4 or 16 circuits on a 4-pole machine or 6, 12, or 24 circuits on a 12-pole machine. Multicircuit wave windings with the same number of circuits as poles can be made by using the same slot and bar combinations as on a lap winding.

For example, with an 8-pole machine with 100 slots and 200 commutator bars, the bar throw for a simplex lap winding would be from bar 1 to bar 2 and then from bar 2 to bar 3, etc. For an 8-circuit wave winding, the winding must fail to close by circuits/2 bars, or 4. Thus, the throw would be bar 1 to 50, to bar 99, to bar 148, etc.

The throw is (bars # circuits/2)(p/2), in this case, (200 # 4)/4 # 49. Theoretically such windings require no equalizers, but better results are obtained if they are used. Since both lap and multiple wave windings can be wound in the same slot and bar combination simultaneously, this is done by making each winding of half-size conductors.

This combination resembles a frog’s leg and is called by that name. It needs no equalizers but requires more insulation space in the slots and is seldom used.

Some wave windings require dead coils. For instance, a large 10-pole machine may have a circle of rotor punchings made of five segments to avoid variation in reluctance as the rotor passes under the five pairs of poles.

To avoid dissimilar slot arrangements in the segments, the total number of slots must be divisible by the number of segments, or 5 in this case. This requires the number of commutator bars to be also a multiple k of 5.

However, the bar throw for a simplex wave winding must be an integer and equal to (bars # 1)(p/2). Obviously (5k # 1)/5 cannot meet this requirement.

Consequently one coil, called a dead coil, will not be connected into the winding, and its ends will be taped up to insulate it completely. No bar will be provided for it, and thus the bar throw will be an integer. Dead coils should be avoided because they impair commutation.

MULTIPLE OR LAP WINDINGS OF DC GENERATORS BASIC INFORMATION

Figure 8-12 shows a lap-winding coil. The conductors shown on the left side lie in the top side of the rotor slot. Those on the right side lie in the bottom half of another slot approximately one pole pitch away. At any instant the sides are under adjacent poles, and voltages induced in the two sides are additive.

Other coil sides fill the remaining portions of the slots. The coil leads are connected to the commutator segments, and this also connects the coils to form the armature winding. This is shown in Fig. 8-13.

The pole faces are slightly shorter than the rotor core.

Almost all medium and large dc machines use simplex lap windings in which the number of parallel paths in the armature winding equals the number of main poles. This permits the current per path to be low enough to allow reasonable-sized conductors in the coils.

Windings. Representations of dc windings are necessarily complicated. Figure 8-14 shows the lap winding corresponding to the Gramme-ring winding. Unfortunately, the nonproductive end portions are emphasized in such diagrams, and the long, useful portions of the coils in the core slots are shown as radial lines.

Conductors in the upper layers are shown as full lines, and those in the lower layers as dotted lines. The inside end connections are those connected to the commutator bars. For convenience, the brushes are shown inside the commutator.

Note that both windings have the same number of useful conductors but that the Gramme-ring winding requires twice the number of actual conductors and twice the number of commutator bars. Figure 8-15 shows a 6-pole simplex lap winding. Study of this reveals the six parallel paths between the positive and negative terminals. The three positive brushes are connected outside the machine by a copper ring T# and the negative brushes by T#.

The two sides of a lap coil may be full pitch (exactly a pole pitch apart), but most machines use a short pitch (less than a pole pitch apart), with the coil throw one-half slot pitch less than a pole pitch. This is done to improve commutation.

Equalizers. As shown in Fig. 8-15, the parallel paths of the armature circuit lie under different poles, and any differences in flux from the poles cause different voltages to be generated in the various paths. Flux differences can be caused by unequal air gaps, by a different number of turns on the main-pole field coils, or by different reluctances in the iron circuits.

With different voltages in the paths paralleled by the brushes, currents will flow to equalize the  voltages. These currents must pass through the brushes and may cause sparking, additional losses, and heating. The variation in pole flux is minimized by careful manufacture but cannot be entirely avoided.

To reduce such currents to a minimum, copper connections are used to short-circuit points on the paralleled paths that are supposed to be at the same voltage. Such points would be exactly two pole pitches apart in a lap winding.

Thus in a 6-pole simplex lap winding, each point in the armature circuit will have two other points that should be at its exact potential. For these points to be accessible, the number of commutator bars and the number of slots must be a multiple of the number of poles divided by 2.

These short-circuited rings are called “equalizers.” Alternating currents flow through them instead of the brushes. The direction of flow is such that the weak poles are magnetized and the strong poles are weakened. Usually, one coil in about 30% of the slots is equalized. The crosssectional area of an equalizer is 20% to 40% that of the armature conductor.

Involute necks, or connections, to each commutator bar from conductors two pole pitches apart give 100% equalization but are troublesome because of inertia and creepage insulation problems. Figure 8-15 shows the equalizing connections behind the commutator connections. Normally they are located at the rear coil extensions, and so they are more accessible and less subject to carbon-brush dust problems.

STRANDING AND TRANSPOSITION OF AC GENERATORS ARMATURE WINDINGS BASIC AND TUTORIALS

A wide variety of winding types may be used to produce a desired voltage with the desired number of phases and a suitable waveshape. In small generators, “scramble wound” armature windings may be used.

However, in most alternator applications, double-layer, form-wound coils in open slots with 60° phase belts are used. In such a winding, each slot has two conductor bars (often called halfcoils), not necessarily from the same phase winding.

These bars are insulated from ground and secured in the slot, usually by wedges. It is usually necessary for the bar to have the ability to slide axially in the slot to accommodate thermal expansion, but it must not be loose in either the radial or azimuthal directions. This has led to a number of proprietary techniques for armature construction.

Winding Forms
Figure 7-20 shows an example winding diagram. For the purposes of this figure, the machine is shown “rolled out flat,” with the dotted lines on either side representing the same azimuthal location. In this case, the machine has 24 slots, each with two half-coils, as shown in the slot allocation section of the drawing, at the bottom of the figure.

FIGURE 7-20 Armature in 24 slots, 5/16 pitch.

The upper part of the figure shows how one phase of the winding would be laid out. This drawing shows a lap type winding (the most commonly used) with a 5/6 pitch. In a 24-slot, 2-pole winding a full-pitch coil would span 12 slots, while in the 5/6 pitch winding the coils span 10 slots.

Fractional Slot Windings.
Fractional slot windings, in which the number of slots per pole per phase is not an integer, have coil groups that differ from one another. These can be arranged to produce balanced voltages under circumstances that are beyond the scope of this discussion.

Stranding and Transposition
At power frequencies (50 or 60 Hz), the skin depth in copper is on the order of 1 cm so that it is usually necessary to subdivide armature conductors into a number of parallel strands. In form-wound coils, these strands are usually rectangular to allow for good space factor.

To prevent circulating currents between parallel strands, it is necessary to employ transposition to ensure that voltages induced in each strand are approximately the same.

The simplest form of transposition, often used in transformers and sometimes in generators, is to twist the armature conductors at 180° in the end turns. Or sometimes, groups of conductors are connected together in the end turns with a progressive transposition that constitutes a “twist” of the winding from half-coil to half-coil.

Transposition of strands in the end turns is generally not satisfactory in large ac generators. A transposition scheme attributed to Roebel is usually used (see Fig. 7-21). The Roebel transposition is equivalent to a twist of the conductors in the slot. It is usual to carry out the Roebel transposition only within the slot part of the winding.

FIGURE 7-21 Illustration of Roebel transposition: (a) typical offset conductor strand; (b) group of conductor strands composing half the conductor; (c) complementary group; (d) assembly.

A variety of transpositions are used, including 180°, 360°, and 540°. The first two are effective only at eliminating circulating currents due to flux crossing the slot, but do not compensate for flux in the end windings.

The 540° transposition, attributed to Ringland and Rosenberg (1959) is often applied because it filters out most of the circulating currents in a bar. Other more complex Roebel transposition arrangements that more extensively filter out circulating currents are possible but are rarely used because of manufacturing complexity.