Eye On Electronics

Motor, Nov 2007 by Dale, Mike

Nonmoving, noncontact sensors have been developed for torque measurement Among possible applications, these new sensors allow improvements in transmission shift control strategies.

Torque measurement is one of the most important variables to be considered in the operation of a vehicle. While torque must be measured at various locations in the vehicle, one area of particular interest is in the transmission. It's known that shift quality could be improved if there were an accurate, instantaneous and cost-effective torque measurement system available. Until recent developments in a new technology called magnetoelasticity, this goal has been elusive.

Most Popular Articles in Autos: Service Slants; 2007 utility vehicle buyer's guide: Side-By-Sides are popular; here's who ...; Transmission considerations: beyond the manual gearbox; Buell Motorcycle engineering, innovation, & dedication: in an industry ...; 100 + 10: America's oldest automotive magazine celebrates its 110th year ...; More »

Understanding torque measurement begins with a clear understanding of torque itself. By definition, torque is rotary motion around a fixed spot or axis. Unlike horsepower, torque does not include die element of time or of distance traveled. Static torque is a force that does not actually create motion. A good example is a bolt that has already been tightened to spec-say, 40 ft.-lbs. Any force up to that amount is called static torque; once the bolt does start to move again, the applied force is called dynamic torque.

Torque sensors fit into two broad categories-reaction and rotary. Reaction-type sensors use fixed, stationary components that do not rotate with the part. Rotary-type sensors use transducers that do rotate with the body to which the torque is applied.

A wide range of basic technologies can be used to measure torque. Which one is used in a given application depends on die accuracy and speed of measurement desired. Like most sensors today, die net measurement is a matter of hardware, software and signal conditioning, including amplification and, in some cases, analog-to-digital conversion. Most types of torque measurements are temperature-sensitive, and the output has to be compensated for this.

An SAE paper written by engineers from Chrysler, Methode Electronics and Magnetoelastic Devices points out some of die difficulties involved in measuring torque. It says, in part: "The determination of when to shift gears, and the control of the complex series of events that perform the shift, requires a variety of sensor inputs and a stored knowledge map of the engine's speed-torque and tbrottle position characteristics. The problem of achieving a good shift is complicated by variations in the engine map with engine temperature, barometric pressure (altitude), engine wear and variations between individual engines. Many investigations have been made in trying to estimate torque from indirect sources such as computations involving engine and torque converter maps and rate of change in shaft rotational speed." To date, none of these efforts has been totally satisfactory.

The primary principle behind the type or torque measurement discussed in this paper is the fact that the shaft bearing the torque load twists in proportion to the load. The twisting has the effect of slightly increasing the diameter of the shaft, while at the same time causing it to become slightly shorter. The torque stress also changes the magnetic characteristics of the shaft. These two phenomena have resulted in many different torque-sensing concepts.

One way to measure the sheer stress the torque creates in the shaft is to use strain gauges that are soldered or glued to the shaft. There are at least three problems with shaftmounted sensors: They tend not to be robust, they can be labor-intensive to install and there is the issue of how to get the torque information from the rotating sensor to the nonrotating part of the vehicle. A further problem with strain gauges glued to the shaft is that there's an rpm limit that if exceeded will result in the sensor being thrown from the surface. Omega Corp., a supplier of torque-measuring equipment, says that reaction-type strain gauges and sensors have the disadvantage of not being completely accurate because they ignore the inertia of the motor.

The SAE paper mentioned considers other possible methods of torque measurement. The torsion bar concept, widely used in torque wrenches, was deemed not accurate enough. Power steering system torque sensors work on the variable reluctance concept. The thought there was that this concept wouldn't work with the shafts that had the stiffness needed for transmissions.

Torque sensing based on magnetic principles avoids many of the other sensor concept pitfalls. One of the primary advantages is that magnetic coupling of sensors to the shaft requires no direct connection to the shaft. This eliminates the possible need for slip rings, sensor elements or other materials necessary to form the signal and get it to the nonmoving parts of the system that need the information.

Magnetostriction is a very interesting magnetic phenomenon first noticed by the scientist James Joule back in the 1840s. He observed that the shape of a sample of nickel he was working with changed when he applied a magnetic field to it. Changing shape as a result of an applied magnetic field is called the magnetostrictive effect. It turns out that this also works in the reverse direction: Changing the shape of the metal-in this case by applying torque-causes changes in the magnetic fields of the shaft. This is called the Vilhry effect. A special aspect of this is that under magnetic fields of a certain intensity, there can be a change in the polarity from positive to negative. This aspect is called the Villary reversal.

Featured Download

Speak Like a CEO

Automotive Industry

Eye On Electronics

Motor, Nov 2007 by Dale, Mike

Find Research Guides for:

Find Featured Titles for: Arts

Most Popular Articles in Autos

Most Popular Publications in Autos