Overrange, A Crucial Rotary Torquemeter Parameter

Rotary power producing and consuming devices produce (or absorb) torque in a pulsating rather than a smooth manner. That’s because they have discrete poles and/or pistons and/or gear meshes, etc., which generate complex forcing functions. Furthermore, drivelines consist of several inertias and torsion springs which resonate¹ at one or more frequencies. As a result, even under “constant load,” real-world driveline torque is never constant. Instead, it consists of an average torque with one or more superimposed torsional components. 

A torquemeter must therefore have an acceptable overrange rating to be error-free over its entire operating range. If not, when installed in a driveline, significant errors can occur. A torquemeter’s overrange is the highest torque at which measurement error is less than 0.1% of full scale. A leading torquemeter manufacturer expresses overrange as a percentage² of full scale. Note: Overrange of an mV/V output torquemeter is the lesser of its overload rating and the overrange of its external signal conditioner. Unless it is adequate, large measurement errors will occur.

Figures 1 and 2 show the response of two torquemeters, each with 0.1% combined error and torquemeter range selected for “best accuracy.” Both torquemeters have a full-scale rating equal to the engine’s rated (average) torque, the generally accepted condition for assuring best accuracy. Figure 1 torquemeter has 150% overrange; Figure 2 torquemeter has no overrange. Both monitor the same driveline, which has a torsional resonance at 45 Hz (2,700 rpm) with peak amplitude equal to half the average torque.

The torquemeter with 150% of full-scale overrange measures peak and average torques without error. The torquemeter without overrange has a 16% error for average torque and a 33.3% error for peak torque. That device incorrectly reports the engine power and related driveline stresses to be much lower than actual.

Substantial, unacceptable real-world errors occur despite a torquemeter combined error specification of 0.1% and with the range selected for best accuracy. The magnitude of such errors is a function of system overrange, the amplitude of dynamic torque perturbations, and the average torque level.

Table 1 illustrates the effect of overrange and dynamic torque peak when the average torque is at full scale, the generally accepted condition for best accuracy. Note, the no overrange device has an unacceptable error even with a dynamic torque peak as low as 10% of full scale.

Many commercial torquemeters, signal conditioners, and most computer data acquisition cards have no or little overrange. Table 2 shows their overrange errors. As noted, such devices operating at full scale have large errors even with small dynamic torques. Unless operated downscale, a no-overrange device has much larger errors when higher dynamic peak torques are present. To avoid such errors, you can greatly oversize the torquemeter. Oversizing increases temperature, linearity, and hysteresis errors, and has other detrimental effects. Adequate overrange, rather than torquemeter oversizing, is the solution.

A torquemeter's overload rating is the maximum torque that can be applied without yielding its element or otherwise producing a permanent change in its performance. Overload is usually specified as a percentage of full scale. Some torquemeters have overload ratings between 200% and 1,000% of full scale.

Repeated torque cycles can cause a fatigue failure when torque peaks are less than a torquemeter's overload rating. Some torquemeters on the market have an infinite fatigue life when subjected to full torque reversals equal to half their overload rating; see Table 3. Thus, when cyclical torques are expected in the overload region, select³ a torquemeter with a 400% (or higher) overload rating. There is a qualified torquemeter for virtually any level of torsional oscillations.

Overrange of premium digital torquemeters is 150 to 300% of full scale, depending on model. Most dc-operated analog torquemeters have 133%  of full-scale overrange. Some instruments have at least 150% of full-scale overrange on their digital output. Their analog outputs have 164% of full-scale overrange on the 5 V setting and 135% of full-scale overrange on the 10 V setting.

In summary, unless a torquemeter’s overrange is specified and is adequate for the driveline dynamic torques, significant real-world errors can occur no matter how impressive its other specifications are. These errors are virtually certain when you use a torquemeter, signal conditioner or data acquisition cCard with no or little overrange.

Footnotes

1. Technical Memorandum 8150, Avoiding the Destructive Effects of Torsional Resonance, S. Himmelstein and Company.
2. The long accepted industry standard for Transducer Percent Overload mathematically is [Torque Overload Value]*100/[Full Scale Torque]. Himmelstein has elected to make the Overrange and Overload definitions consistent; i.e., define Overrange Percentage as [Overrange Torque]*100/[Full Scale Torque]. To further improve clarity, new Torquemeter Specifications and reprints of older Himmelstein specifications will also contain actual Overrange Torque in engineering units.
3. Bulletin 705, Choosing the Right Torque Sensor, S. Himmelstein and Company.