Over the last 40 years, permanent-magnet couplings have come into the manufacturing forefront. This technology can operate at high speeds without contact or wear of active components, and at temperatures from cryogenic to 500° F. Permanent magnetic couplings also correct misalignments and provide overload protection within a single set of components. Once applied only to chemical pumps, magnetic couplings are now used in a variety of applications, ranging from downhole oilwells, liquid nitrogen pumps, high-pressure compressors, severely misaligned drivelines, and non-contact vibration isolation drives for encoders and torque transducers.

Let's compare the three basic methods of mechanical power transmission using permanent magnets.

  1. Synchronous power

    In this arrangement, one magnet follows another. Two sets of strong magnets are arranged on opposing components. The pole pieces attract one another so that the magnet sets stay in synch and torque is transmitted. There is no relative motion between the driver and driven — only static rotation relative to the pole pieces due to torque. When the torque rating of this coupling/clutch is exceeded, the two parts slip and very little torque is transmitted. Depending on rotational velocity, the parts reengage at the next set of poles. Magnets are arranged on two coaxial cylinders, one attached to the drive and the other to the driven. Alternately, magnets can be attached on two opposing plates.

  2. Hysteresis power

    Like with synchronous power, in this arrangement one magnet follows another. Here, a strongly magnetic driver is attracted to a weaker magnetically driven material. Torque is transmitted until it reaches the magnitude at which the weak material slips. When this happens, transmitted torque holds constant as the weak material is magnetized and demagnetized by close proximity of the strong magnet, and the two parts move relative to one another. In some cases, substantial heat is generated. Then once the overload is removed, the components automatically reengage.

    Slip torque can be adjusted by changing the relative position of opposing strong magnets on either side of the weak magnetic material, or by varying the magnet gap or overlap location.

  3. Eddy current

    In this case, Lenz's Law (Vinduced = -∂Φ/∂t) and the properties of conductors in a magnetic field are leveraged. To review, Lenz's Law states that induced current flows to electromagnetically oppose the motion or cause by which its produced.

A series of rotating magnets is held in close proximity to a highly conductive material, such as aluminum or copper. The rotating magnetic field sets up opposing electrical eddy-current fields in the conductive disc, causing them to produce a magnetic field that drags against the opposing rotating magnets. Eventually, the conductive disk nearly catches up with the driving magnets.

Unless a lockup clutch is used, there is always slippage in eddy current units. This type can be constructed with variable torque by adjusting the phase or air gap of the magnets. Due to generation of the eddy currents in the conductive disc, adequate cooling is necessary to prevent overheating. This cooling is normally supplied by airflow during rotation; however, at lower speeds external cooling is required.

Leveraging nonmechanical linkage

Since there is no physical contact between driver and driven parts, magnetic couplings yield many interesting possibilities of completely isolating the driving and driven components; in fact, driven components can be totally separated from the drive by solid walls. This is most frequently incorporated with synchronous units but can be applied to other designs as well.

Because the interaction between parts is magnetic instead of physical, the resulting forces on driving and driven components is much less than conventional connections — and complete torsional and electrical isolation are possible.

However, not all barriers are created equal. When designing couplings and clutches to transmit rotation across solid barriers, the effects of Lenz's Law are significant. In metallic barriers, rotating magnetic fields can induce electrical eddy currents and their corresponding magnetic fields. This loss — in the form of heat and drag on driving components — is calculated:

where

K = a constant depending on design

T =thickness of barrier

L = length of magnets

N = speed, rpm

Bg = flux density of the magnets

D = mean diameter

M = number of sets of magnets

R = electrical resistivity, μOhms per cm3

This equation shows that losses can be considerable at higher speeds. Additionally, for a given geometry, the higher the specific electrical resistance of the barrier, or the lower the magnetic field, the lower the losses. From a practical standpoint, grade-five titanium, Hastelloy, nickel alloys, 300-series stainless steels, beryllium copper, and other nonmagnetic alloys can be used as barriers. However, aluminum must be avoided due to its higher conductivity. With aluminum, the effects of Lenz's Law are apparent even at hand-operated speeds.

Since losses are proportional to flux density B2 excess magnetic power is not beneficial and should be avoided. To decrease field strength across the barrier as much as possible, the opposing magnets need to be displaced from one another as much as possible. At maximum torque, magnets are not directly across from one another; instead they are displaced in torsional rotation nearly to the midpoints between magnets; this yields minimum field through the barrier.

Low safety factors and soft starts are used to avoid oversizing the parts. The heat generated by eddy currents in the barrier must be removed with cooling fluids to avoid overheating of the barrier. At lower speeds or magnetic field torques, the amount of heat generated may be removed by conduction and convection to the environment without additional cooling circuits. Alternatively, a non-metallic barrier (for strength, a composite or ceramic) can be used to eliminate eddy currents and respective heat and losses.

Hysteresis for constant torque

In units that operate by hysteresis, the amount of magnetic field penetrating the weaker magnetic material can be varied by constructing the coupling with stronger magnets on opposite sides of the weaker material. The strong magnets are arranged such that one set can be moved relative to the other.

When like poles face each other, they produce maximum magnetic saturation, forcing lines of flux to travel circumferentially through the hysteresis disc for maximum torque. When opposite poles face each other they produce minimum saturation of the hysteresis disc, since the lines of flux travel right through it. A combination of adjustment angles between these two extremes gives infinite adjustability.

Alternately, one or two sets of strong magnets can be moved away from the weaker material (for less overlap or increased air gap) to vary the amount of torque transmitted. Because there are no contacting surfaces, the setting can be maintained indefinitely. Once it is set, this torque does not change. Continuous slip within operational parameters do heat the clutch but does not change its torque or necessitate resetting as with wearing friction, wrap spring, or ball detent types.

Hysteresis products are normally made with magnets that keep their properties constant over the expected temperature range. However, hysteresis units sometimes need additional cooling to avoid overheating other components during extended periods of overload and slip.

Other things to note

In synchronous and hysteresis drives, coaxial versions operate in axial equilibrium. In other words, when properly installed the inner and outer magnets float axially relative to each other and tend to align themselves. There is no axial force induced on driving or driven components. However, coaxial designs are not stable radially; any radial misalignment causes inner and outer components to attract one another. For this reason, both the inner and outer drives must be fully supported.

On the other hand, disc styles are very stable radially (generating relatively low radial reaction forces) and can transmit a significant percentage of full load torque even when grossly radially misaligned. The two discs are strongly attracted to one another and must be axially restrained against coming together under normal operation. When overloaded, the like poles are opposite and push away from one another; they should be restrained in the opposite direction. This potential axial motion can be used to trip a limit switch on overload. One caveat: Allowing axial motion of one of the hubs does reduce the torque rating of the coupling/clutch, because the increased air gap lowers slip torque.

Magnetic couplings are like most other mechanical power transmission products: A larger diameter makes for higher torque. Too, in coaxial models, increasing the length results in a nearly linear increase in torque. This is important where large pressure differentials (even to 20,000 psi) use small-diameter couplings with relatively thin walls. In these situations, low magnetic field strength is used. In the extreme case, high power (400 hp at 3,600 rpm, for example) can be transmitted across a solid barrier — by say, a five-inch-diameter coupling to over six feet long.

For more information, visit magnetictech.com or e-mail the editor at eeitel@penton.com.
Courtesy Magnadrive Corp.

Eddy-current drives

Eddy-current drives can also transmit variable torque by changing the distance between the conductors and drive magnets. Many people have used eddy-current drives without knowing it; automobile speedometers are eddy current drives with a return spring. Exercise bikes that use eddy-current drives are also quite common; as the user pedals faster, drag increases proportionally. On eddy-current drives, slip is constant under all conditions and the effects of Lenz's law is proportional to the square of speed. These products are typically used at speeds above 1,000 rpm, and may need forced air cooling to operate at lower speeds.

A word about magnet types

Most magnetic couplings and clutches are now made with high-energy rare-earth magnets. These fall into two categories: NdFeB (neodymium iron boron, also called simply neo) and SmCo (samarium cobalt.) SmCo has undergone improvements in recent years for increased power density. SmCo5 designs are being replaced with Sm2Co17.

All magnets have some reversible field losses as well as irreversible losses with increasing temperature. At even higher temperatures (Curie temperatures) they lose all magnetic properties, and the magnetic fields become totally random. On a cubic-inch basis, NdFeB is stronger than SmCo for similar degrees of magnetization at room temperature. However, as temperature increases the NdFeB magnets become weaker faster than SmCo — 0.1% per °C versus 0.03% per °C. For NdFeB magnets, irreversible losses may occur starting at 120°F (for standard grades) vs. 150° C for high-temperature UH and EH grades. Usually for applications over 200° F, SmCo is used for its more constant torque rating over all temperatures.

Some older products and lower power density designs (in other words, those requiring larger volume) include ceramic or AlNiCo. These magnets are less expensive per cubic inch. However, it must be noted that AlNiCo magnets (made from an alloy of aluminum nickel and cobalt) used in the very earliest synchronous couplings are prone to demagnetization and must be carefully applied.