A REVIEW OF STEEL MATERIALS IN MOTION DEVICES

George P. Gogue
Joseph J. Stupak, Jr.

G2 Consulting,
Beaverton, OR 97005

I) Introduction:

Electrical steel is the name given to the steels used in motion devices like motors and actuators. Major developments over the past century have given us a wide selection of new steel alloys while constantly improving on the old ones. The general trend is driven by variety, economy and efficient devices.

Following is a discussion of some of the most common parameters and features associated with electrical steel which are crucial in developing the ability to choose the appropriate material for the particular device.

II) Classification of steels:

There are many ways to classify electrical steels. Each contributes to one or more of the performance parameters of the final device. They are also of varying importance to the final application in view of non-technical issues such as economy, availability and manufacturability. Following is a list of the classification criteria:

1. Grade

2. Orientation

3. Processing

4. Production

5. Composition

6. Thickness

7. Surface

1. Grade designation:

The grades of electrical steel were devised in accordance with the core losses of the materials under certain conditions. (These conditions were typically 15,000 G and 60 Hz on a sample of .014 inch thickness.) The most widely used classification is by the American Iron and Steel Institute (AISI) where the materials are designated with the letter M followed by a number.

The number used to refer to the core losses (in W/lb) under the above mentioned conditions but has become obsolete in view of the advancement in steel materials.

2. Orientation:

This refers to the orientation of the material grains. The material is called non-oriented (motor grade) if it shows the same magnetic characteristics in any direction of magnetization (in the plane of the material). However, by a special manufacturing process, a material can be made to exhibit a preferred magnetic performance in a particular direction of magnetization.

This preference for one direction of magnetization also results in poor performance in all other directions. Performance criteria will be explained in later sections. Table 1 lists the steels according to orientation and grade.

3. Processing:

The oriented materials listed in Table 1 are produced then annealed by the manufacturer after production. They are, therefore, called fully processed. However, the non-oriented steels in Table 1 may or may not be annealed by the manufacturer prior to shipping. If they are not annealed they are referred to as semi-processed. In that case, it becomes the responsibility of the user to anneal the material to achieve the full performance. The effect of annealing on performance will be discussed in more detail in Section VII.

4. Production:

The final sheets of steel are manufactured by a process of rolling slabs of electrical steel. Initially the process is done at high temperature producing hot-rolled steels in coil form. After some chemical processing, these coils are subjected to further rolling at room temperature to achieve the final thickness. The steel is then called cold-rolled steel. This final operation improves the mechanical characteristics of the steel without affecting the magnetic characteristics.

5. Composition:

The superior magnetic characteristics of electrical steel are attributed mainly to the small amounts of silicon added to the alloy. These alloys have a higher resistivity than those without silicon, resulting in lower eddy-current losses (as will be seen later). Silicon addition results in lower hysteresis losses in non-oriented steel by virtue of its effect on the grain structure. As the silicon content of the steel is increased, the hardness and yield strength increase. However, the brittleness increases and the ductility decreases as the silicon content increases.

Table 2 shows the chemical composition of typical silicon electrical steels. Other metals and, in particular, aluminum and manganese have a favorable effect on the grain structure, resulting in lower hysteresis losses. The other elements present in the steel are considered impurities and are reduced to the low levels given in the table by the manufacturing process.

6. Thickness:

Steel in various thicknesses can now be obtained from manufacturers. However, certain thicknesses are more available due to their wide use in electrical devices. These values correspond to gages known as the Electrical Steel Standard Gage (ESSG). Table 3 lists these gages, with the most popular being denoted by *. These gages have a nominal thickness given by ESSG which corresponds to the range given by the Manufacturers' Standard Gage (MSG).

Oriented steel is usually also available in thicknesses smaller than those listed in Table 3. The density of electrical steels depends on the silicon content of the material and not on the gage. The American Society of Testing Materials (ASTM) recommends appropriate values of density for these alloys as given in Table 4.

7. Surface:

The surface finish of electrical steel can either be the natural oxide film that forms in the manufacturing process or one of several types of coating. The choice of surface coating is important to the gains from laminating the steel being realized, as will be explained in Section IV.2 in the discussion on eddy-current losses. Table 5 lists some of the most common types and their corresponding applications.

Figure 1: BH curves

Notes to Figure

1. The curves shown are intrinsic. The flux-densities shown represent the increase in field strength caused by the material, not the total field. To get the normal curve (used for design), add the value of B found from the curve to the value of (H times 1 Gauss/Oersted). For example, for Vanadium-permandur at 1000 Oe the curve shows 22,600 G. The flux-density in an immediately adjacent gap would be 23,600 G.

2. The data shown is plotted from information from various sources, and is of unknown accuracy. It is given for illustrative purposes only. Design data should be obtained from the material manufacturer.

3. The materials are shown in their highest state of permeability. Some alloys require special heat-treatment to reach their best magnetic properties, and may be adversely affected if subsequent machining and forming is done. The 3% silicon-iron sheet is anisotropic, and the best magnetic properties are obtained in only one direction.

III) Magnetization curves:

The magnetization curve of a material describes the relationship between the 2 most important parameters of that material, i.e., B & H. The flux-density or induction B is a measure of the concentration of the magnetic field in the material (units in kG or Tesla). The magnetizing force H is the force necessary to create the field in the material (units in kOe or AT/m). The ratio of B/H is called the permeability () and is a measure of how easy it is to magnetize a particular material.

Figure 1 shows magnetization curves of typical magnetic materials. These are called soft materials because B collapses to a very small value upon the removal of the magnetizing force H. This value of B is called remanence or residual flux-density . If, however, is a significant portion of the saturation level and the material becomes a permanent magnet, it is then known as a hard material.

A typical magnetization curve can be divided into 2 sections. The first section is at low levels of H and it stretches to the "knee" of the curve where the second section starts. As the magnetizing force H gets higher and higher, the material goes further into saturation with smaller and smaller gains in the flux-density B. The saturation level for electrical steels depends on the silicon content of the material. However, the range is small compared to the saturation level of other alloys as shown in Figure 2.

Curves of the permeability () are shown in Figure 3 for a few materials besides the silicon steels. In motion devices, it is generally desirable to use these materials in the range of highest permeability. This results in the highest B for a fixed amount of H. The values in Figure 3 are equal to total B divided by total H. Values of the differential permeability, however, are calculated from the ratio of change in B to that in H. The differential permeability is actually a description of the continuously changing slope of the BH curve.

IV) Core losses:

The core losses are the losses experienced by the electrical steel as a result of the same varying magnetic field it is intended to carry. These losses appear in motion-devices as heat, limiting the maximum operating temperature of the device. The two main constituents of these losses are the hysteresis and eddy-current losses.

1. Hysteresis loss:

The hysteresis loss in steel is a result of the change in the direction of flux in that steel and its non-linear magnetization behavior. The latter is explained by the peculiar shape of what is called the hysteresis loop shown in Figure 4.

Point a is at the saturation level of the steel referred to in Section III. Reference was also made there to the residual magnetism Br remaining in the steel after the magnetizing force is reduced to zero.

In fact, a certain amount of magnetizing force is required in the reverse direction to bring the flux-density to zero at point b. A further increase of in the reverse direction would take the flux-density to saturation at point c. This is numerically equivalent to point a but in the third quadrant. The path from c to a follows the same steps as from a to c.

As a further illustration, Figure 5 shows an electromagnetic fixture having a magnetomotive force caused by a current-carrying coil with a variable gap. If the gap is permitted to close completely, the loop reluctance becomes very low and the magnetic flux-density, B, increases to a high value. The coercivity, H, in the material also increases. If the coil current is now cut off, that source of mmf vanishes but the material coercivity, H, remains. Although H may be small, the loop reluctance is so small that a significant amount of flux still links the circuit. Then the pole pieces are held together by high magnetic forces and the circuit will not open. This unintended magnetic effect is called remanence. The magnetic field may gradually dissipate over a period of time ranging from a tenth of a second up to thirty seconds. This is apparently due to such causes as small field variations, nearby electrical equipment, lights, wires, thermal atomic motion etc. One good remedy for remanence in a magnetic circuit is to ensure by some means that the gap never closes to the degree necessary to maintain significant coercivity in the pole material.

A boundary between two well-machined steel surfaces which are clamped together, still offers some impedance to magnetic flux. For surfaces of about 64 micro inches surface roughness, which is a fairly good finish in steel, the boundary behaves magnetically like an equivalent air gap of perhaps .002 inch. If the surfaces are polished or lapped, however, the equivalent air gap is much smaller, and a magnetic circuit with such surfaces is much more susceptible to remanence effects.

The hysteresis energy loss of the material is proportional to the area enclosed by the path abcda. The hysteresis power loss in W/lb can then be determined by multiplying by the frequency f of the flux reversals. The following expression combines these variables.

where

: hysteresis loss (W/lb)

A : loop area (gauss-oersted)

f : frequency (Hz)

: steel density (gm/cc)

For a particular steel material the hysteresis loss can be calculated for all values of saturation flux-density . These values have been shown to vary approximately with the 1.6 power of for values of between 500 G and 15 kG. This factor is called the Steinmetz coefficient.

2. Eddy-current loss:

If electrical steel is present in the vicinity of an electrical coil, flux will flow in the steel as a result of the alternating current in the coil. The flux in the steel will itself induce an emf in the material following the basic laws of induction. Since the material is essentially an electrical circuit closed on itself, the induced emf will cause a circulating electrical current called an eddy-current. Its value is dependent on the value of emf and on the resistivity of the path of current. As in any other electrical circuit the power loss is the product of the square of the current times the resistance. In a similar manner to the hysteresis loss, the eddy-current loss manifests itself as heat, contributing to the maximum operating temperature limit of the device.

From the above discussion, it would seem that a reduction of the eddy-current losses can be achieved by reducing the emf induced in the material and by increasing the cross-sectional area of the flux-path. Breaking a solid piece of steel into several laminations divides up the total flux among all the laminations. The emf induced in each lamination would be equally reduced and so would the resulting eddy-current. A reduction in the eddy-current is due to the higher resistance of the restricted flux-path. Since power varies with the square of current, using laminations results in appreciable reduction in eddy-current losses.

Eddy-currents flowing in steel give rise to a magnetic field which opposes the original flux. This interaction between the two fluxes results in a non-uniform distribution of the flux in the steel and leads to what is called "skin effect". This phenomenon effectively reduces the cross-sectional area of the flux-path. Using lamination instead of solid steel corrects this problem by allowing the use of the largest possible path area.

It is, therefore, advantageous to use the thinnest possible lamination especially in high frequency application. This is determined by manufacturing limitations and economic considerations. It was shown (by Steinmetz) that the eddy-current loss is proportional to the square of the lamination thickness.

The following equation shows the variables affecting the eddy-current loss.

where

: eddy-current loss (W/lb)

: constant (0.5 to 1)

t : thickness (inch)

f : frequency (Hz)

B : flux-density (kG)

: resistivity (micro Ohm-cm)

In addition to the slicing of a solid piece of steel into laminations to reduce the eddy-current losses, it is essential to ensure physical separation between the laminations. In Table 5, various classes of coating were listed enabling a suitable selection to fit the application. Even the thinnest coating of the natural oxide can substantially reduce the eddy-current losses as compared to the non-coated stack of lamination.

As mentioned above, the amount of eddy-currents flowing in the steel is dependent on the resistivity of the path of current. Therefore, by appropriate selection of a steel alloy with the highest possible resistivity, the minimum possible eddy-current is realized. Increased silicon content of the electrical steel linearly increases the resistivity of the material. Figure 6 shows the resistivity owing to differing silicon contents as well as various other materials to the same scale. These values are only true at a room temperature of 20°C.

There are, however, some disadvantages in increasing the silicon content of electrical steel. The most noticeable is a drop of approximately 500 G in the saturation level for each percentage point increase in Si. In other words, the permeability at high levels of induction is reduced as a consequence of a larger amount of silicon content.

A less important disadvantage of increasing the silicon content of the electrical steel is the associated drop in the Curie temperature. Over the usual range of silicon percentage content, a drop of approximately 17°F is observed for each percentage point increase in the silicon content. This is not of significant importance unless the device using this steel is expected to operate at very high temperatures.

3. Total losses:

The loss data for electrical steels given by manufacturers are usually of the total core losses and are a combination of hysteresis and eddy-current losses. Only in exceptional cases are these losses given separately.

Core-loss data is usually available in curve form over the useful range of flux-density. It is sometimes possible to find data on apparent core losses as well as the usual core losses. The difference being in the former referring to the losses in RMS volt-amp/lb whereas the latter are in W/lb.

Typical conditions for which the core losses are given are for 60 Hz and a particular thickness gage. Table 6 gives the values of these losses for 3 popular thickness gages.

V) Stress & Strain:

Two types of strain result from the application of force (stress) on electrical steel or from shearing and punching. The application of a steady low force results in elastic strain. This strain is lost after the force on the steel is removed. However, after shearing or punching the steel, plastic strain is introduced and is retained even after the removal of that stress. The effect of these strains, especially the plastic, can be quite substantial on the magnetic characteristics of the steel. This is essentially so because of the distortion in the pattern of crystals in the material. The BH and loss curves, normally supplied by the manufacturer for a particular material, do not hold true any more when those strains are present. By one estimate, the percentage increase in hysteresis loss resulting from shearing a 1.5 inch wide lamination was equivalent to using one thicker grade of steel than the one sheared.

Plastic strain can also be introduced in the steel as a result of bending, either during storage or when a coil is straightened in preparation for use. In general, any deformation of the steel from the condition during processing by the manufacturer results in changes to the magnetic performance of the material. This affects both oriented and non-oriented steels.

VI) Magnetostriction:

When a magnetically permeable part is placed in a magnetic field, dimensional changes occur in the part. Nickel contracts in the direction of magnetization and expands in the transverse directions. Permalloy, on the other hand, reacts in the opposite way, expanding in the direction of magnetization and contracting in other directions. The following chart gives change in length (inches per inch of length) at magnetic saturation for various materials, in the direction of magnetization:

Magnetostriction may cause magnetic assemblies to radiate considerable noise. Since the material changes length twice (once for the positive and once for the negative lobe) for each cycle of a sine wave, the primary sound frequency generated is twice that of the exciting electric current for a reversing magnetic field. If the structure experiencing magnetostriction is mechanically restrained, stresses are developed in the material which may in turn lead to such effects as a large decrease in permeability and saturation field strength.

VII) Annealing:

To return the material to its original state, electrical steels are annealed in one of many procedures. Strains are relieved by annealing and the material is returned to a stress-free condition. Annealing is initially performed by the manufacturer on fully-processed steel as was mentioned in Section 11.3. However, the process must be repeated if the steel was subjected to stresses in the process of uncoiling, shearing and stacking.

The annealing process allows the crystals to regain their original orientation if they have been distorted and enables crystal growth at the edges where shearing creates stress lines. Relieving the stresses caused by stacking laminated steel to make cores can be done by annealing of the complete core after stacking.

The process of annealing essentially involves elevating the temperature of the steel at a particular rate to a top value and holding it at that level if necessary before allowing it to return to room temperature. The whole cycle is carried out in a controlled atmosphere to avoid chemical contamination of the steel. In particular, oxygen and carbon compounds (such as those present in lubricants) should be avoided. Water vapor can result in oxidizing the surface of the steel and in increasing the material thickness. The recommended atmosphere is dry nitrogen with 5 to 10% hydrogen.

In general, the annealing process should be as short as possible, without causing a thermal shock to the material. The temperature gradient should take into account the mass and size of the material being annealed to prevent warping of laminations and to ensure uniform heating of the surfaces. This gradient can vary widely, e.g. 40 to 300°F/hour.

The maximum temperature needed to return stressed steel to stress-free condition is usually 1500 +/- 50°F. The minimum soak time requirement at that temperature should be sufficient to allow heat to travel to the innermost parts of the stack.

The cooling process can also be done in the shortest possible time and at the same gradient as the heating process. The gradual cooling can be terminated at about 600°F and the steel removed from the protective atmosphere without any bad consequences.

VIII) Magnetic alloys:

Other alloys of steel which do not contain silicon are also used occasionally in motion devices. These magnetic alloys may have one or more parameters which are significantly better than those of silicon steel but usually at the expense of other parameters. Their chemical composition usually results in higher cost compared to silicon steel. The following Table 7 lists the composition and properties of some of these materials. The balance of material composition beyond the values given in the table is naturally iron. 3% silicon oriented steel is also listed in the table for comparison.

The high permeability of Nickel iron alloys makes them ideal for shielding magnetic field. The low saturation flux-density of these materials, however, is immaterial in those applications. The 50% Nickel alloy, however, has a moderate saturation flux-density making it suitable for some motion devices. These devices would benefit from the low core losses of this material.

Vanadium permendur, on the other hand is suitable for very high flux-densities but has fairly low values of permeability.

The relatively recent addition to the selection of steels is Metglas. The major advantage of this material is in the extremely low power loss and high permeability. This material is presently available in ribbon form only and, therefore, is only suitable for the construction of special devices.

Alloys of steel with 12% or more of chrome have improved resistance to corrosion and are often called "stainless" steels. There are three main groups of stainless steels, the austenitic, ferritic and martinsitic. Austenitic stainless steels, e.g. types 302, 303, 304, 310, 316 and 324 contain 6% or more of nickel as well as chrome. They are nonmagnetic in their annealed state and cannot be hardened by heat treatment. They can be hardened and strengthened by severe cold working (e.g. cold forging of bolts) and some of these steels then become magnetic.

Ferritic and martinsitic stainless steels are designated by 400 Series numbers. A few of these contain relatively low amounts of nickel (up to 2.5%); most contain none. Both types are magnetic. The ferritic stainless steels, e.g. types 405, 406, 430 and 446 cannot be significantly hardened. The martinsitic stainless steels, e.g. 403, 410, 431 and 440 can be hardened by heat treatment. Type 440C is often used in the hardened state for roller ways, bearings and instruments. The 400 Series stainless steels are less corrosion-resistant than the 300 Series steel. Some 400 Series steels may form rust from condensed moisture in a warm and damp environment.

Certain iron alloys containing about 30% nickel have permeabilities which are strongly temperature-dependent. One of these alloys (Carpenter's 30 Type 2) has a relative permeability of 130 at -40°F, which decreases to only 20 at 120°F (at 46 Oersteds). These materials may be used as magnetic flux shunts to compensate for magnetic flux variations with temperature in critical applications.

References:

1. "Electrical Materials Handbook", Allegheny Ludlum Steel Corp., Pittsburgh, PA, 1961

2. "Selection of Electrical Steels for Magnetic Cores", Armco Inc., Middletown, OH, 1985

3. "Non oriented Sheet Steel for Magnetic Applications", United States Steel, Pittsburgh, PA

4. "Carpenter Soft Magnetic Alloys", Carpenter Technology Corp., Reading, PA, 1974

5. "Motor Laminations", Tempel, Niles, IL

6. "Metglas", Allied Chemical, Parsippany, NJ