Log in
  
Home > Sensors > Force/Strain/Load/Torque
Related topics: Force/Strain/Load/Torque,Aerospace, Powertrain, Sensors Mag, Technologies in Depth
Force/Strain/Load/Torque

Steps Toward High-Precision Torque Measurement

June 1, 2006 By: Bob Davis, HBM Sensors

Here are some guidelines for selecting and mounting a torque flange—and a new digital-type flange—that could save you time and headaches.


Rising fuel prices, coupled with tightening regulations on vehicle efficiency and emissions, are creating a global demand for higher levels of precision in torque measurement on test stands for engines and transmissions. On one hand, torque values must be certifiable to meet regulations and manufacturers' specs. On the other, they must be precise enough to pick up efficiency differences of 0.1% or less in the test object. Such tight system accuracy requirements mean that the torque flanges themselves must be accurate to within 0.05%.

Item: Some automakers are developing minimal-lubricant transmissions and units that use expensive synthetic fluids as part of an effort to reduce both friction and hydrocarbon content in the vehicle—while attaining 99.3% efficiency. This calls for further improvement of measurement uncertainties. That level is well beyond the capabilities of conventional slip-ring mechanical torque sensors and most digital types in the real world.

Item: Pressure to streamline manufacturing throughout the automotive and turbine industries is creating a demand for more compact, more versatile test stands that also deliver high accuracy. This just about rules out older barrel-type torque sensors because of their size, mass, and dynamic range limitations.

Fortunately, newer digital torque flanges are able to meet these demands. We'll take a look at one example (Figure 1) later on. But let's first consider some general principles of high-precision torque measurement.

Figure 1. The new T12 digital torque flange
Figure 1. The new T12 digital torque flange

Some Basic Principles

Know What You Need. The main problem is that even with a "perfect" torque sensor the measured value can be no better than the basic design of the test system in its entirety, the care taken to mount the sensor, and the appreciation of all sources of system error. That is to say, you need to understand exactly what you want to know and what the data you collect really mean. You could be measuring the wrong thing or drawing the wrong conclusions from the data. Or you could connect the torque flange the wrong way for the measurement you are seeking.

Don't Overspecify. Define reasonable requirements for system accuracy. This is important strategically as well as technically. At the design stage for an engine, turbine, or transmission, more accurate torque data can give you a distinct advantage in a very competitive world market. The more accurate the data, the closer to the limits you can safely design. Torque differences due to different lubricants in an engine, though small, are nonetheless important in the global auto business. And it takes a sensitive torque test to reveal them.

For these reasons, developmental torque testing of engines and drivetrains may increasingly need 0.3% system accuracies, attainable only with properly installed advanced digital torque flanges. Calibration may require 0.008% total system accuracy. By contrast, most production cold-engine testing in automotive plants needs only 1%–3% system accuracy.

Overspecifying accuracy or resolution requirements will add unnecessary cost, slow down the test, consume memory, and actually make it harder to find what you are looking for. It's the same principle as sending images over email. A 1000 dpi image takes twice as long to print or send as a 500 dpi image, twice the file space to store, and twice the time to retrieve or scan.

Recently, for instance, a turbine manufacturer inquired about components for a production test stand, insisting on total system accuracy within 0.1% over the entire dynamic range. This simply isn't possible because of the physics of the system. Errors arise from too many sources: the coupling shaft, the test-stand transmission, temperature variations, structural deflection, and sensor mounting errors, to name just a few. After an hour's discussion, he still wasn't satisfied and said he'd call other providers. A week later, he called us back with an order for what we had recommended in the first place. In the process, he lost a week from his schedule and a lot of man hours. 

1 2 3






Sensors.tv