TORQUE

Conventional, or Elastic, Methods of Torque Sensing

The vast majority of conventional systems used to measure torque operate by measuring the torsional deflection induced by the applied torque, by either of two methods:

a. Measurement of the twist angle

The twist angle method of torque measurement generally requires a slender portion of the shaft to enhance the twist (several degrees at most for L/D = 5) and a pair of identical toothed disks attached at opposite ends of the slender portion. The twist angle can be determined from the phase difference between magnetically or optically detected tooth/space patterns on each of the disks. This method generally requires the shaft to be rotating.

b. Measurement of the surface strain

Changes in surface strain can be measured by piezoresistive strain gages attached to the shaft. These strains are generally too small (at most a few parts of 103) to be accurately measured directly. Common practice is, therefore, to use four gages arranged in a Wheatstone bridge circuit. With rotating shafts, coupling means, such as rotary transformers, are required to feed the excitation current to the gages and to acquire the signal from the bridge circuit in a non-contacting manner.

In both of these elastic systems, the high compliance required to obtain accurately measurable deflections in the normal operating torque range must be taken into consideration. In addition, the supporting electronics tend to be relatively large and complex. Many of these disadvantages reflect the fact that at the most basic level, elastic torque sensors produce signals that are a function of torsional strain. This fact alone compromises their ability to be applied to shafts of arbitrary torsional stiffness, and to offer a frequency response above 1kHz.

Brief Introduction to Magnetoelastic Torque Sensors

Magnetoelastic torque sensors produce signals that are a function of torsional stress, not strain. As a result they are generally much stiffer mechanically than the conventional elastic torque sensors. Additionally, they offer significantly higher frequency response, typically on the order of 2-4kHz. Indeed the determination of surface stress from the measurement of magnetic quantities by magnetoelastic methods provides an inherently non-contacting basis for measuring torque in a more compact construction than those required for either the twist angle or surface strain elastic methods. Magnetoelastic torque sensors measure magnetic quantities related to the surface shear stress by either of two methods:

a. PB Type I - "Permeability Based"

In Type I, the permeability changes in the shaft surface, caused by the stress-induced magnetic anisotropy, affect the permeance of a magnetic flux path which includes a magnetizing source and a pickup (sensing) coil.

b. PB Type II - "Polarized Band"

In Type II, the stress-induced magnetic anisotropy causes a remanently magnetized magnetoelastically active member to generate a measurable magnetic flux.

Traditional, PB Type I "Permeability Based", Magnetoelastic Torque Sensors

As explained above, in a Type I torque sensor the output signal derives from changes in permeabilities of regions on the shaft surface due to the magnetic anisotropy that arises with transmitted torque. Specifically, larger values of permeability occur along the stress induced easy axis and smaller values occur along the hard axis which is perpendicular to the former. Advantages of the Type I permeability-based magnetoelastic torque sensor stem from its naturally wireless transduction and mechanically robust construction. Various specific excitation/pickup constructions can be utilized, but local variations in magnetic properties of typical shaft surfaces limit the attainable accuracy in both the branch and the cross pickup constructions. In order to eliminate many of the rotational issues associated with the cross and branch Type I constructions, another important group of Type I torque sensors was developed utilizing an axisymmetric solenoidal construction. In this design, oppositely directed helical grooves are machined or formed (typically along ±45° angles to the axis) on adjacent circumferential regions of a steel shaft, and solenoidal coils encircling these regions are used for excitation and sensing. The axial permeability of a grooved region increases when the easy axis of the stress-induced magnetic anisotropy occurs in parallel to the line of grooves, whereas it decreases otherwise. This results in different voltages being induced in the sense windings encircling the two regions and this difference provides the measure of the torque. The solenoidal construction for Type I magnetoelastic torque sensors, by way of the axisymmetric structure of the windings, hides local variation of the magnetic properties of the shaft.

Despite their various benefits, Type I permeability-based magnetoelastic torque sensors suffer from a number of deficiencies which have limited their proliferation in real-world applications. These deficiencies ultimately stem from the fact that the variable being measured, permeability, does not depend exclusively on the applied torque. In particular, it is important to note the following:

• Permeability is not an intrinsic property of a magnetic material
• Permeability is not a single-valued, structure-sensitive property
• In any one material composition, processed in one controlled manner, permeability will still generally vary over a large range both with temperature and with magnetization
• Permeability will also generally vary with frequency

The end result is that the range of permeability variation with most of these factors, which do come into play in many practical applications, often exceeds the changes in permeability that are a function of torque, which is the quantity of interest.

MagCanica's PB Type II, "Polarized Band" Torque Sensor Technology

MagCanica is a leader in the development of the novel Type II, or polarized band, class of magnetoelastic torque sensors. These sensors provide many of Type I sensors' benefits of noncontact, robust, compact construction, while at the same time overcoming many of their problems. In MagCanica's PB Type II, or Polarized Band sensors, magnetization of the active region does not occur continuously in service but rather is carried out one single time before the sensor enters service. Sensor operation is based on the reorienting effects of torsional stress on the individual magnetic moments that have been remanently circularly magnetized. In response to the magnetoelastic energy associated with the biaxial principal stresses by which torque is transmitted along the shaft, each moment will rotate towards the nearest positive principal stress direction and away from the nearest negative principal direction. This reorientation of the originally circular magnetization results in a net axial magnetization component. The divergence of this component at the edges of the polarized bands is the source of a magnetic field in the space around the shaft which can be readily measured with one or more magnetic field sensors.

Type II torque sensors can be constructed either with a thin ring of magnetoelastically active material rigidly attached to the shaft, or by using a portion of the shaft itself as the magnetoelastically active element. Subsequently, when torque is applied, the magnetizations tilt into helical directions, causing magnetic poles to develop at the central domain wall and at the end surfaces. The polarity of the magnetic poles reverses when the applied torque changes its direction, and so does the output signal accordingly. Torque is determined by measuring magnetic flux with one or more magnetic field sensors. Notice that permeability variations do not come into play in the Type II sensor, for the quantity being measured is an externally detectable magnetic field whose intensity is linearly proportional to the shear stress (and by extension, the applied torque).

Schematic diagram of the operating principle underlying MagCanica's PB Type II Polarized Band torque sensor system.

This concept can be readily applied, for example, to a practical application such as a driveshaft. Polarized bands are applied to a dedicated measurement region on the shaft, and an array of field sensors (FS) is placed in two diametrically opposed pairs around the shaft circumference to detect the field created by the shaft under torque.
Schematic diagram of a driveshaft having two polarized bands, surrounded by an array of magnetic field sensors (FS) to detect the magnetic field under torque.

Schematic diagram of a driveshaft having two polarized bands, surrounded by an array of magnetic field sensors (FS) to detect the magnetic field under torque.

Benefits of MagCanica's Torque Sensor System

The following is a list of the key characteristics of MagCanica's magnetoelastic, non-contact torque sensor products. These characteristics represent our core advantages that differentiate our products and services from those of our competitors:

• Increased torsional stiffness
• Reduced size and mass
• Non-invasive measurement
• Packaging flexibility
• Reduced complexity
• Naturally wireless mode of measurement
• True non-contact solution
• Outstanding frequency response and dynamic torque measurement (up to 4.5 kHz)
• Scaleable technology (from less than 1 Nm to over 200,000 Nm)
• Applicable to various production aerospace shaft materials
• Numerous HUMS (Health and Usage Monitoring) and Prognostic Health Monitoring (PHM) applications

[Back to the top of this page]

RATE-of-CHANGE of TORQUE (ROC)

Introduction

Beyond the nominal torques being transmitted which can be measured accurately with MagCanica's torque sensor system, torque variations are ubiquitous in the rotating parts of machines and can be measured with MagCanica's rate-of-change of torque (ROC) sensor system. Sensing and measurement of the variational torque components can often provide more detailed information concerning machine function than might be apparent from a measurement of torque alone, which is often dominated by its larger, more steady state components. This benefit is analogous to that obtained by using an accelerometer in an airbag sensor system to detect an impending crash, as opposed to a vehicle speed sensor which would not be able to detect a collision rapidly enough in the first place, and sometimes not at all due to the insufficient sensitivity to rapid velocity changes that an accelerometer is able to detect but a vehicle speed sensor is not.

The Need for ROC Measurement

Torque variation can be fully characterized by its time rate-of-change (ROC). Continuous measurement of ROC of torque can provide sufficient information to identify specific torque producing events, recognize "signature" abnormalities, and signal the need for controlling action. While ROC information can, in principle, be obtained by differentiating a torque dependent signal, such an approach would fundamentally limit signal quality. MagCanica's ROC sensor system overcomes these issues to reliably detect several forms of time-varying torque that can be of interest such as:

• Torque variations reflective of the actual function of the machine (e.g., impulse wrenches, rock crushers, etc.)
• Torque variations originating from kinematic features (e.g., oscillating or reciprocating parts)
• Major torque variations originating from inconstant rates of energy input or usage (e.g., as in piston engines, air compressors, etc.)
• Torque variations that vary with rotational position such as those arising in electric motors (e.g. cogging in machines having permanent magnet rotors or ripple in brushless motors)

Some sources of variational torque, such as those which arise on, or from the rotation of, ship propellers, helicopter rotors or wind turbine rotors, are more subtle in that they arise simply from the variation with rotational position of the blade's proximity to the hull or fuselage or frame, respectively. Moreover, all such active variational torques can stimulate torque oscillations at frequencies dependent on the dynamic relationship between inertia and elasticity within individual or interconnected rotating parts.

MagCanica's ROC Sensor System - Architecture and Benefits

MagCanica's ROC system has been successfully implemented on such varied applications as automotive engines and transmissions, electric motors, and machine tools. The system operates according to the measurement chain shown in the diagram below:

Signal Flow Diagram for MagCanica's rate-of-change of torque (ROC) sensor system

MagCanica's ROC system provides a number of complementary benefits beyond those provided by the torque sensor system:

• The frequency response of an ROC system is much greater than that of any existing torque sensor system
• The detectable amplitude of a given variational torque is higher for higher frequencies in an ROC signal than that which can be practically obtained through a differentiated torque sensor signal
• Small amplitude, rapid changes in torque representing potentially very important diagnostic or control data, are either not easily observable by measuring torque alone or not at all, whereas they are detectable by measuring ROC directly
• Fewer restrictions on types of acceptable shaft steels
• The ROC system can withstand very high temperatures, even in excess of 300ºC

[Back to the top of this page]

DRAG FORCE METHOD (DFM) FOR NON-DESTRUCTIVE EVALUATION

Conventional Magnetic Flux Leakage(MFL) Methods of NDE

Conventional Magnetic Flux Leakage (MFL) methods of non-destructive evaluation (NDE) are based on the fields which arise from the divergence of magnetization in inhomogeneous or physically discontinuous material. The inhomogeneity can be associated with either (or both) dimensional or permeability variation. It is easy to visualize that fields will arise in the vicinity of those "flaws" which can be characterized as "missing" material, e.g., holes, dents, corrosion, scratches, foreign matter inclusion, and the like. Conventional MFL techniques all rely on magnetizing local portions of the sample being examined (using permanent magnets or electromagnets), and sensing the fields in proximate space. Interpreting the signature features of these fields provides information on the location, size and other characteristics of the causative flaw. Conventional MFL techniques generally utilize Hall Effect field sensors, or pick-up coils when there is continuous motion of the sample or magnet

MagCanica's Drag Force Method (DFM) of NDE

MagCanica's Drag Force Method (DFM) of non-destructive evaluation constitutes a significant innovation and offers a number of potential benefits. It employs force sensors to sense the interactive force between a permanent magnet (which does the magnetizing) and the dipole moments which arise with inhomogeneous magnetization of the underlying material being scanned for defects. If the material is homogeneous and free of defects, these forces will be equal and opposite (i.e., symmetrical) on either side of the magnet's neutral plane. While the material will be thereby stressed, there will be no net tangential (drag) force acting on either the material or the magnet. Inhomogeneity such as that caused by a crack, for example, clearly upsets this balance, and the resulting net force is sensed by the force sensor. Note that the force measured in the Drag Force Method is not the normal attractive force which acts to bring the material being scanned and the magnet closer together. In fact, the material being scanned and the magnet must each be supported in such fashion as to prevent motion in this direction.

Schematic diagram of the Drag Force Method (DFM). The permanent magnet (PM) scans over the sample under test (SUT), generating a net tangential drag force indicative of the level of underlying inhomogeneity in the sample, such as that which might be caused, for example, by a crack or a void.

Benefits of MagCanica's DFM and Development Plans

The key benefits of the Drag Force Method, i.e. sensing the reaction force on the magnet, compared with the conventional Magnetic Flux Leakage method, i.e. sensing the field associated with the flaws, are the following:

• Conventional Magnetic Flux Leakage NDE systems utilize field sensors, especially galvanomagnetic (Hall effect or magnetoresistance) sensors, that are small and have highly localized sensing regions, and therefore require the use of substantial numbers (10-100) in arrays to simultaneously scan significant areas. Such an arrangement inherently implies significant complexity and processing capability.
• Drag Force Method NDE systems utilize a magnet/force sensor combination that can scan an area an order of magnitude larger than the equivalent magnet/field sensor combination that would be used in an equivalent MFL system

MagCanica is in the early stages of development of the DFM, and is now beginning to explore its range of applications. As part of this effort, MagCanica is developing a variety of magnet/force sensor arrangements, each configured for a specific application. First-generation DFM devices currently use a pendulum-mounted magnet and a strain gauge load cell.

[Back to the top of this page]

EMERGING TECHNOLOGIES

In addition to the core torque sensor technology, the rate-of-change (ROC) technology, and the drag force method (DFM) technology, MagCanica is also pursuing various other development programs in the following areas expected to lead to new sensing devices:

• Combined torque-and-bending measurement
• Axial position measurement
• Speed measurement

[Back to the top of this page]