Conversion Formulas Potential Measurement Errors What is involved in building a Primary Torque Standard Proper Handling techniques

Transcription

1 FUNDEMENTALS OF TORQUE CALIBRATION 1

2 Lesson Objectives Introduction Definition and Examples of Torque Conversion Formulas Torque Calibration Traceability Standard Types of Torque Standards ASTM E2428 BS7882 Accuracy and Precision Performance and Uncertainty Potential Measurement Errors What is involved in building a Primary Torque Standard SECTION 2 IN CLASS DEMONSTRATION ONLY (NOT ON WEB) Torque Wrench types Proper Handling techniques Torque Wrench Demonstration 2

3 Company History 1920 s Morehouse and the U.S. Bureau of Standards started to design and refine force calibration products (Proving Rings) for the purpose of generating an accurate force for Brinell Hardness Testing. Pictured above: Morehouse Brinell Proving Ring S/N 14 Calibrated by U.S. Bureau of Standards test # May 24, 1926 (86 years ago!) 3

4 Morehouse Proving Ring S/N 14 Calibrated in 1926 and the last calibration we have on record is July 25, In 58 years there was a shift of.012% at capacity. 4

5 Company History 1930 s The Morehouse Proving Ring was refined and used to calibrate Material Testing Machines s Morehouse developed products for commercial industry including, Force Gauges, Morehouse Universal Calibrating Machines and Morehouse Dead Weight Primary Standards for calibration of load cells, proving rings etc.., 5

6 Company History 2004 Morehouse becomes the first accredited commercial calibration laboratory to offer dead weight primary standards calibrations accurate to 0.002% of applied force up to 120,000 LBF 2009 Morehouse expands force calibration range offering ASTM E74 calibrations up to 2,250,000 LBF in compression and 1,200,000 LBF in tension. 6

7 Company History Morehouse finishes construction of new torque calibration laboratory. This calibration laboratory features a primary torque calibration standard accurate to 0.002% of applied torque. This standard was acquired from the National Physical Laboratory in England which is a National Metrology Institute Morehouse becomes A2LA Accredited for Torque Calibration 7

8 Calibration Lab Pictures 8

9 General Information 9

10 Morehouse FUNDEMENTALS OF TORQUE CALIBRATION Henry Zumbrun Morehouse Instrument Company 1742 Sixth Ave York, PA PH: web: Sales: 10

11 Morehouse FUNDEMENTALS OF TORQUE CALIBRATION William Lane, P.E. Morehouse Instrument Company 1742 Sixth Ave York, PA PH: web: Technical: 11

12 Contents Torque Definition of Torque Examples of Torque Conversion Formulas Torque Traceability Standard Calibration Traceability Types of Torque Standards ASTM E2428 BS7882 Accuracy and Precision Performance and Uncertainty Uncertainty Distributions Potential Measurement Errors 12

13 Contents (continued) In Class Demonstration Morehouse Torque Calibration Laboratory Conceptual Design Air Bearings Lever and Reaction Beam Boron Tapes Couplings Performance and Uncertainty 13

14 The Importance of Torque Control The object of a threaded fastener is to clamp parts together with a tension greater than the external forces tending to separate them. When the bolt is torqued properly it remains under constant stress and is immune from fatigue. 14

15 The Importance of Torque Control If the torque is not applied properly and the tension on the bolt torque is too low, varying loads will act upon the bolt and it will fail. 15

16 The Importance of Torque Control Honda Recalling 50,000 Vehicles July 25,2011 TOKYO (Kyodo) -- Honda Motor Co. said Monday it will recall a combined 50,122 units in the Stream minivan, the Civic sedan and the Crossroad sport utility vehicle to repair their bolts that fix the water pump pulley to the engine, free of charge. In a report filed with the Ministry of Land, Infrastructure, Transport and Tourism, Honda said the engines of the vehicles, made between July 2008 and July 2010, may stall due to the defect. The automaker said it will also implement a recall for the models in overseas markets, including Latin America and Europe, where a total of 150,000 units were marketed. The ministry said the bolts may lose tension or fracture, causing the drive belt which runs both the alternator and the water pump to circulate coolant water to come off the pulley. 16

17 The Importance of Torque Control If the tension is too high, the tightening process may cause bolt failure. Pictured Above: Metal snap from Jeff Nihel s dragster, apparently the bolts on the left exhaust manifold were over-torqued bolts then failed, manifold popped off and 4000bhp of exhaust gas launches the car in the air at over 200mph! 17

18 The Importance of Torque Control Imagine if one of the one hundred and fifty plus car engine bolts is undertorqued, it loosens overtime and eventually destroys the engine. What if the bolts are under-torqued in an airplane assembly and become loose in midflight? Fastener reliability depends on controlling the tightening torque. Other engineering factors such as fastener material, design, pitch & surface finish may also influence the tightening torque. 18

19 Torque Is derived and traceable to SI units of length, mass, and time. The meter is the length of the path travelled by light in vacuum during a time interval of 1/ of a second. The kilogram is the unit of mass; it is equal to the mass of the international prototype of the kilogram. The second is the duration of periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium 133 atom. 19

20 Torque = Force x Length CIPM/BIPM defines 1N as the force required to accelerate one kg to one meter per second per second in a vacuum. Most torque systems consist of an arm or disc of a known length and a set of weights. 20

21 Torque = Force x Length Local Gravity - Determination of gravity must be done at the site performing the calibration and the elevation of each weight should be considered. (Varies by up to.2% across the U.S.) Material Density - Knowing the density of the material from which the weights were manufactured. Buoyancy the Archimedes principle applies, ie. air pressure under the weights causes an upwards force. This reduces the effective force generated by the weights and therefore the mass must be increased to allow for this. (When used in air the correction is approximately 0.015%) 21

22 Definition of Torque Torque, also called moment or moment of force is the tendency of a force to rotate an object about an axis, fulcrum, or pivot. Just as a force is a push or a pull, a torque can be thought of as a twist. Torque (BS 7882 definition)- product of tangential force and length applied about a known center of rotation 22

23 Definition of Torque Loosely speaking, torque is a measure of the turning force on an object such as a bolt or a flywheel. For example, pushing or pulling the handle of a wrench connected to a nut or bolt produces a torque (turning force) that loosens or tightens the nut or bolt. 23

24 Examples of Torque 24

25 Examples of Torque (cont) 25

26 Torque = Force x Length Example 1: Distance = 1m. Force = 2000 N, Torque = 2000 N-m Example 2: Distance = 2m. Force = 1000 N, Torque = 2000 N-m Example 3: Distance = 1 ft. Force = 1000 LBF, Torque = 1000 FT-lbf. Example 3: Distance = 5 ft. Force = 200 LBF, Torque = 1000 FT-lbf. 26

27 Force Conversion Factors newton (N) to LBF = MULTIPLY BY newton (N) to KGF = MULTIPLY BY newton (N) to ozf = MULTIPLY BY lbf to newton (N) = MULTIPLY BY lbf to KGF = MULTIPLY BY lbf to ozf = MULTIPLY BY.0625 kgf to lbf = MULTIPLY BY kgf to newton (N) = MULTIPLY BY Kgf to ozf = MULTIPLY BY ozf to newton (N) = MULTIPLY BY ozf to lbf = MULTIPLY BY 16 Ozf to kgf = MULTIPLY BY

28 Length Conversion Factors METER (m) to FEET (ft) = MULTIPLY BY METER (m) to INCHES (in) = MULTIPLY BY FEET (ft) to METERS (m) = MULTIPLY BY FEET (ft) to INCHES (in) = MULTIPLY BY INCHES (in) to FEET (ft)= MULTIPLY BY 12 INCHES (in) to METERS (m) = MULTIPLY BY

29 Torque Conversion Formulas Force 1N = lbf Length 1m = ft 1 N-m = x = FT-lbf Force 1N = lbf Length 1m = in 1 N-m = x = in-lbf 29

30 Torque Conversion Formulas If I need to apply 100in-ozf and have a 100mm Calibration Disc. How much weight do I need to apply? Convert 100mm to m 1000mm = 1 m 100mm/1000 =.1m Convert.1m to in =.1 * = in-ozf = in * Force 100 / = approx 25.4 ozf 30

31 Calibration Traceability Guidlines Check the website of accrediting bodies for the accredited laboratories individual scope of accreditation ts.nist.gov/standards/acceditation/index.cfm The 1 st page of any ISO is useless after it has been determined that the laboratory is accredited. 31

32 Calibration Traceability Guidelines It is the scope of accreditation that determines the laboratory s capability. The cert should state the best measurement uncertainty for different ranges. On this cert the best measurement uncertainty is % for torque calibrations from N-m 32

33 Torque Traceability The Torque Machine at Morehouse belonged to a NMI (National Physical Laboratory) This machine was verified in international CIPM test to be accurate to % of applied torque 33

34 Morehouse Primary Torque Standard In 2009 Morehouse purchased the Primary Torque Standard from the National Physical Laboratory (NPL) in England and began construction of a new Torque Calibration Laboratory. Construction of the Laboratory was completed in March On completion of construction the Primary Torque Standard was shipped to Morehouse and NPL sent a team of engineers and scientists to commission the machine. This Primary Torque Standard was used and registered with BIPM as the national standard for the United Kingdom. 34

35 Morehouse Primary Torque Standard The machine was used in a recent CIPM (International Committee for Weights & Measures) international measurement comparison and demonstrated a world leading measurement uncertainty. It is the goal of Morehouse to use this machine to promote US Commerce by dissemination of accurate torque measurements. 35

36 CIPM Comparison 36

37 Types of Torque Standards Lever deadweight This consists of a set of calibrated masses that act on a calibrated lever arm. The system can be directly applied to the transducer in the case of an unsupported calibration beam. 37

38 Types of Torque Standards (cont) In the ideal case the lever arm will be connected to a bearing to eliminate bending from the weights and masses and to minimize friction. This type of system is used for static torque calibration. 38

39 Types of Torque Standards Lever deadweight typically tend to be chain loaded Some consist of multiple weight stacks with various weights or multiple weight stacks of the same size. 39

40 Types of Torque Standards Lever dead weight systems are typically used with two different types of torque arms. Single ended Dual ended / Radius ended Torque is generated by the application of a known force at a known radius from the center of rotation of the torque transducer. Radius ended beams are typically designed with a +/- 8 degree usable arc within which the calibration accuracy is unaffected. Single ended beams do not have a constant overturning moment as the dual radius arm would have and may have additional error as compared with Radius ended beams. 40

41 Types of Torque Standards Lever deadweight with pneumatic or chain loaded weight stack(s). - This system eliminates shock loading associated with manual lifting. This type of system is used for static torque calibration and offers the lowest uncertainties. 41

42 Types of Torque Standards Reference torque transducer Torque can be applied using a motor or hydraulics with the torque controlled by means of a calibrated reference torque transducer in a feedback loop. Other similar designs may take traceability from calibrated load cells measuring the reaction force at the end of the lever arm. 42

43 Types of Torque Standards (cont) This type of system can be used to measure static torque but has the additional advantage of being suitable for continuous torque calibration, whereby the applied torque is applied over a much shorter time. The disadvantage of these systems is that the uncertainty of applied torque will be much higher because the system is dependent on the prior calibration of the reference. 43

44 Uncertainty of Torque Standards Regardless of the torque facility to be used, it is important to evaluate the uncertainty of the system. This should include contributions from all influencing parameters (e.g. mass, length, alignment, and environmental factors). 44

45 ASTM E2428 Torque Calibration ASTM E2428 Definitions Primary torque standards a dead weight force applied through a lever arm or wheel, with a calibrated length or radius of a known uncertainty, that is traceable to national standards. This type of standard is needed to assign Class AA loading ranges. Secondary torque standard an instrument or mechanism, that has been calibrated by a comparison with a primary torque standard(s). This instrument will have a Class AA loading range and may be used to calibrate 45 other instrumentation and assign Class A loading ranges to other instrumentation.

46 ASTM E2428 Selection of Calibration Torque Values At least 10 torque values should be applied to provide an adequate and unbiased sample of the full range of deviations. These values should be selected and distributed at least one calibration torque for every 10% interval. If the transducer is to be used under one tenth of capacity an eleventh point of no less than 400 x the resolution for Class A loading range and 1667 x the resolution for Class AA should be included. 46

47 ASTM E2428 Number of Calibration Torque Values A total of at least 30 torque applications per mode, clockwise or counter clockwise, is required for a calibration and, of these, at least 10 must be at different torque values. Apply each torque value at least twice during the calibration in both the clockwise and counter clockwise direction, as applies. Examples: On a flange 3 runs of 10 or 11 points should be applied On a square drive 4 runs of 10 or 11 points should be applied 47

48 ASTM E2428 Number of Calibration Torque Values 30 +points reduces standard measurement error 48 T distribution = sample vs mean, Z = population

49 ASTM E2428 Procedural Order In Calibration After setting up the transducer and equipment, the maximum torque to be calibrated should be applied to the instrument at least twice. Morehouse recommends applying this torque value 3-4 times to achieve stability in zero torque indication. After the pre-loading, apply torque values in sequential steps as uniformly as possible without inducing shock or vibration. (If you are manually lifting weights of a large capacity it is not acceptable to drop the weights onto the pan) Morehouse has developed dead weight machines to reduce the error associated with handling large weights. It is recommended that no more than five incremental torque values be applied without a return to zero. 49

50 ASTM E2428 Procedural Order In Calibration After data has been recorded for the first calibration run it is necessary to rotate the torque cell before repeating the next series of torque values. If you are using a torque cell with a square drive you should perform up to 4 rotations (40 to 44 total points) to fully characterize the cell. - This is not required by ASTM E2428 but each orientation on a square drive will yield different outputs. Otherwise, the torque transducer should be calibrated in three different positions in each mode it is to be used (clockwise, anti clockwise). 50

51 ASTM E2428 Calibration Data 51

52 BS 7882 Torque Calibration Some BS 7882 Definitions Calibration torque torque with traceability derived from national standards of mass, length and time, and of specified uncertainty of measurement, which can be applied to the torque measuring device. Lower limit of calibration lower value of torque at which a torque measuring device of a given class can be calibrated. Reference standard equipment used to generate or to measure the reference torque applied to the torque measuring device that is being calibrated 52

53 BS 7882 Torque Calibration BS7882 tests for the following Relative error of indication - mean deflection for a given value of increasing torque minus the corresponding value of applied torque. Relative error of interpolation difference between the value of the mean deflection for a given value of increasing torque and the corresponding calculated value of deflection for the given torque, obtained from a mathematically fitted curve. Relative repeatability closeness of the agreement between the results of successive measurements from the same applied torque, carried out under changed conditions of measurement 53

54 BS 7882 Torque Calibration Relative reproducibility - closeness of agreement between the results of successive measurements from the same applied torque, carried out under changed conditions of measurement Relative residual deflection maximum residual deflection obtained from all the series of torques Relative reversibility difference between the deflection obtained from the last given torque series applied in an increasing mode and the deflection obtained from the same given torque applied in a decreasing mode. 54

55 BS 7882 Torque Calibration Residual deflection algebraic difference between the indicator readings before and after the application of a single series of torques Resolution the smallest discernable measurement interval on the torque measuring device indicator Unlike ASTM E2428 several European standards which include DIN (German Standard) and BS7882 (European Standard) classify torque instruments by class ranges. 55

56 BS 7882 Torque Calibration Uncertainty of calibration torques - Values for the maximum permissible uncertainty of the calibration torques applied for the determination of different classifications of the torque measuring devices. 56

57 BS 7882 Calibration Procedure Number of calibration orientations For classes 0.05 and.1 the torque device shall be calibrated in 3 positions each rotated 120 degrees or in 4 positions each rotated 90 degrees about the measurement axis. For all other classes, the torque device should be calibrated at a minimum of 2 different positions. Warm up - it is recommended to allow the system to be energized for at least 15 minutes prior to calibration Preloading procedure - preload should be applied in succession a minimum of 3 times and should be maintained for 1 to 1.5 minutes. 57

58 BS 7882 Calibration Procedure Select a series of at least 5 approximately equally spaced increasing values of torque from 20% to 100% of the maximum applied torque. If the calibration is required to be made below 20% then torque steps of 10%,5% and 2% of the maximum torque may be used. After the preloading procedure apply two series of increasing torques without changing the mounting position (readings should be taken 30 seconds after the torque is applied). 58

59 BS 7882 Calibration Procedure After the first 2 runs have been taken for determination of relative repeatability, remount the torque device 120 degrees or 90 degrees for square drives and repeat incremental forces. The instrument will be remounted and increasing torques will be applied 1 or 2 additional times depending on the drive. Note: If relative reversibility is required, a single series of decreasing values shall be applied at the end of a series of increasing torques 59

60 BS 7882 Calibration Data 60

61 ASTM E2428 VS BS7882 Uncertainty Analysis BS7882 uncertainty analysis includes the standard uncertainty associated with the calibration torque and tests for reproducibility, repeatability, residual deflection and reversibility. ASTME2428 uncertainty analysis tests for reproducibility with rotation. 61

62 Accuracy and Precision It is a common mistake to assume that an accurate device is precise or that a precise device is accurate. 62

63 Accuracy accuracy is not the same as uncertainty! I Can have an accurate measurement with a large uncertainty. 63

64 Accuracy Accuracy is a measure of bias (estimate of a systematic measurement error) Accuracy: Closeness of agreement between a measured quantity value and a true quantity value of a measurand. Accuracy does not tell you about the quality of the instrument. 64

65 Accuracy Example Example: A torque cell was calibrated against a primary torque standard and at N-m the recorded output on the indicator was N-m. This cell was determined to be accurate to +.1% of full scale or +.5 N-m. Note: from this example the resolution of the instrument may be.1 N-m or.5 N-m and although the resolution may be the same as the accuracy in this example they are different. 65

66 Accuracy Example A picture in which we hope the accuracy of the time remaining is wrong 66

67 Resolution Resolution - the smallest amount of input signal change that the instrument can detect reliably Resolution - is the ability of the measurement system to detect and faithfully indicate small changes in the characteristic of the measurement result. 67

68 Precision Precision is a measure of spread Precision refers to the repeatability of measurement. 68

69 Precision Example Example: 500 N-m was applied 3 times using a primary torque standard to a torque cell and the torque cell indicator s recorded output was N-m, N-m, N-m. Conclusion: The instrument is precise to +/-.01 N-m when 500 N-m is applied. 69

70 Accuracy and Precision A torque cell that may be accurate in one orientation may not repeat when rotated. A torque cell may repeat when a torque is applied, but may not read close to the desired engineering units value. 70

71 Accuracy and Precision 71

72 Uncertainty a measurement is useless unless there is some way of estimating the associated uncertainty good analyses should include uncertainty estimates 72

73 Uncertainty Statistical Uncertainty is defined as the estimated amount or percentage by which an observed or calculated value may differ from the true value. In Metrology, two common types of uncertainty evaluation are Type A and Type B. 73

74 Performance & Uncertainty Type A Uncertainty - method of evaluation of uncertainty by the statistical analysis of series of observations Repeatability condition of measurement, Inter-mediate precision condition of measurement, Reproducibility condition of measurement 74

75 Performance & Uncertainty ASTM E2428 calibration tests for the reproducibility condition of measurement and is an example of Type A Uncertainty 75

76 Performance & Uncertainty Type A Example A series of measurements are taken to determine the uncertainty of the measurement. (Standard Deviation) 76

77 Performance & Uncertainty Type B Uncertainty - method of evaluation of uncertainty by means other than the statistical analysis of series of observations. Type B uncertainty - Evaluation based on information associated with a quantity value of a certified reference material, - obtained from a calibration certificate or manufactures specifications, obtained from the accuracy class of a verified measuring instrument, obtained from limits deduced from a test or experiment. Examples include torque cell temperature effect, drift, resolution, etc.., 77

78 Performance & Uncertainty Type B evaluation method: The method of evaluation of uncertainty of measurement by means other than statistical analysis of a series of observations Examples: Based on specification History of parameter Other knowledge or test of the process parameter 78

79 Performance & Uncertainty Type B Example The temperature effect on torque cell output is +/-.004 % per degrees Celsius The specification of the torque arm is +/ inches 79

80 Uncertainty Distributions The next example will be dealing with two types of Uncertainty Distributions. Normal and Rectangular 80

81 Uncertainty Distributions Normal Distribution Normal distribution Normal distribution is one way to evaluate uncertainty contributors so that they can be quantified and budgeted for. Normal Distribution helps understand the magnitude of different uncertainty factors and understand what is important. The normal distribution is used when there is a better probability of finding values closer to the mean value than further away from it, and one is comfortable in estimating the width of the variation by estimating a certain number of standard deviations. 81

82 Uncertainty Distributions Rectangular Distribution Rectangular distribution is the most conservative distribution. The manufacturer has an idea of the variation limits, but little idea as to the distribution of the uncertainty contributors between these limits. It is often used when information is derived from calibration certificates and manufacturer s specifications. 82

83 Performance & Uncertainty for Calibration performed by Morehouse Type A and B uncertainty analysis COMPANY TORQUE CELL MANUFACTURER TORQUE CELL S/N CAPACITY ASTM E2428 Uncertainty for K=2 THE LOWEST FORCE AT WHICH THE SECONDARY STANDARD WILL BE USED PRIMARY TORQUECALIBRATION STANDARD UNCERTAINTY K=1 PRIMARY ELECTRICAL CALIBRATION STANDARD UNCERTAINTY K=1 (IF APPLICABLE) CAL DATE CALCULATED VALUES TORQUE CELL UNCERTAINTY IN % FOR FULL SCALE K= % TORQUE CELL UNCERTAINTY IN % FOR FULL SCALE K= % TORQUE CELL UNCERTAINTY IN % FOR LOWEST FORCE APPLIED K = % TORQUE CELL UNCERTAINTY IN % FOR LOWEST FORCE APPLIED K = % Type A Uncertainty % MOREHOUSE MOREHOUSE SAMPLE 1000 N-m N-m 100 N-m 0.001% 10/27/2010 Uncertainty Description Uncertainty Distribution Divisor Standard Uncertainty Squared ASTM E2428 Uncertainty % at the lowest calibration torque to be used % normal E E-08 Combined Type A Uncertainty 1.10E E-08 Type B Uncertainty % Uncertainty Description Uncertainty Distribution Divisor Standard Uncertainty Squared PRIMARY TORQUE CALIBRATION STANDARD UNCERTAINTY 0.001% rectangular E E-11 PRIMARY ELECTRICAL CALIBRATION STANDARD UNCERTAINTY (IF APPLICABLE) 0.000% rectangular E E+00 STABILITY OF THE SECONDARY TORQUE STANDARD OVER TIME 0.005% rectangular E E-10 MISALIGNMENT ERROR 0.000% rectangular E E+00 DISSEMINATION ERROR (FOR CALIBRATION LABORATORIES) 0.000% rectangular E E+00 TEMPERATURE ERROR +/- FROM CELL SPEC SHEET % rectangular E E-10 Combined Type B Uncertainty 3.74E E-09 FOR K= 1 UC= SQUARE ROOT OF TOTAL COMBINED TYPE A AND B % FOR K=2 U = K * UC (UNCERTAINTY % AT THE LOWEST FORCE TO BE APPLIED) 0.023% N-m 83 UC AT CAPACITY * 2 ( (UNCERTAINTY % AT INSTRUMENT CAPACITY) 0.008% N-m

84 Performance & Uncertainty for calibration performed with a calibrated beam and masses Type A and B uncertainty analysis COMPANY TORQUE CELL MANUFACTURER TORQUE CELL S/N CAPACITY ASTM E2428 Uncertainty for K=2 THE LOWEST FORCE AT WHICH THE SECONDARY STANDARD WILL BE USED PRIMARY TORQUECALIBRATION STANDARD UNCERTAINTY K=1 PRIMARY ELECTRICAL CALIBRATION STANDARD UNCERTAINTY K=1 (IF APPLICABLE) CAL DATE CALCULATED VALUES TORQUE CELL UNCERTAINTY IN % FOR FULL SCALE K= % TORQUE CELL UNCERTAINTY IN % FOR FULL SCALE K= % TORQUE CELL UNCERTAINTY IN % FOR LOWEST FORCE APPLIED K = % TORQUE CELL UNCERTAINTY IN % FOR LOWEST FORCE APPLIED K = % Type A Uncertainty % MOREHOUSE MOREHOUSE SAMPLE 1000 N-m N-m 100 N-m 0.010% 10/27/2010 Uncertainty Description Uncertainty Distribution Divisor Standard Uncertainty Squared ASTM E2428 Uncertainty % at the lowest calibration torque to be used % normal E E-08 Combined Type A Uncertainty 1.10E E-08 Type B Uncertainty % Uncertainty Description Uncertainty Distribution Divisor Standard Uncertainty Squared PRIMARY TORQUE CALIBRATION STANDARD UNCERTAINTY 0.010% rectangular E E-09 PRIMARY ELECTRICAL CALIBRATION STANDARD UNCERTAINTY (IF APPLICABLE) 0.000% rectangular E E+00 STABILITY OF THE SECONDARY TORQUE STANDARD OVER TIME 0.005% rectangular E E-10 MISALIGNMENT ERROR 0.020% rectangular E E-08 DISSEMINATION ERROR (FOR CALIBRATION LABORATORIES) 0.020% rectangular E E-08 TEMPERATURE ERROR +/- FROM CELL SPEC SHEET % rectangular E E-10 Combined Type B Uncertainty 1.77E E-08 FOR K= 1 UC= SQUARE ROOT OF TOTAL COMBINED TYPE A AND B % FOR K=2 U = K * UC (UNCERTAINTY % AT THE LOWEST FORCE TO BE APPLIED) 0.042% N-m 84 UC AT CAPACITY * 2 ( (UNCERTAINTY % AT INSTRUMENT CAPACITY) 0.035% N-m

85 Performance & Uncertainty The combined uncertainty of a good torque transducer should be approximately 5 times better when calibration is performed using the NPL/Morehouse Primary Torque Standard as compared with a conventional Calibration Beam and Masses. Torque transducers on the Market today are capable of uncertainties much lower than that attainable by primary standards as defined by ASTM E

86 Torque Calibration 2+2 Joke So the teacher assigns to Ada, Bob, Charles and Danna to go home and figure out what is Ada, the daughter of a mathematician, asks her dad. He responds: "Well, = = = 4, but it can be rewritten as 2 + 2, so = 4 Bob asks his mom, who is an engineer. She takes out her HP calculator, punches in RPN the appropiate keys, and announces: "It is Charles asks his dad, the phycisist, and he responds: "Well, it is about pi on a zeroth order calculation Finally, Danna ask his dad, who is an accountant: "Dad, how much is 2 + 2?" And he responds: "How much do you want it to be?" 86

87 Torque Calibration There are several ways to measure an instrument and arrive at the same answer, however if you are not verifying your equipment against known standards then the answer can pretty much be whatever you want it to be. 87

88 Problems With Torque Measurement In The United States Accurate Torque Measurements require traceability back to a National Metrology Institute (NMI) or Morehouse Without verification of the torque system by inter comparison with a National Metrology Institute or Morehouse some of the Type B uncertainty error in the torque calibration system will be unknown. 88

89 Problems with Torque Measurement In The United States Generating an accurate torque is more complex than having a N.I.S.T traceable calibrated beam and N.I.S.T traceable calibrated masses. A comparison study done by the American Society for Testing and Material ASTM Committee E28, Calibrations Sub-committee between several torque calibration laboratories, revealed several laboratories were operating outside of their uncertainty budgets, proving that inaccuracies are occurring throughout North America. 89

90 Problems with Torque Measurement In The United States PICTURED ABOVE A NORBAR TORQUE WRENCH CALIBRATOR. THIS SYSTEM SUPPORTS THE TORQUE WRENCH AS TO NOT INTRODUCE ADDITIONAL HUMAN ERROR. IT IS IMPORTANT THAT THE TRANSDUCERS IN THIS SYSTEM ARE CALIBRATED BY A LABORATORY FOLLOWING THE PROPER CALIBRATION PROCEDURES TO ENSURE THAT THE TRANSFER FROM THE TRANSDUCER TO THE TORQUE WRENCH IS ACCURATE. 90

91 Potential Measurement Errors Using mass weights instead of force weights Introduction of bending and parasitic effects Misalignment Calibrated beam deflection and temperature Drives on the beam or transducer being worn Bending forces on an unsupported beam are unavoidable 91

92 Using Mass Instead of Force Weights It is very important that the gravitational value for the Laboratory is established. The effect of not doing this could be a variation in the force produced by the weight of perhaps 0.25% or more of reading. It is therefore strongly recommended that you establish the local value of gravity (g) for your Laboratory and use weights that have been calibrated at that gravitational constant. The ideal solution is to have the gravity measured on site by the national geological survey agency. 92

93 Weight Handling The physical lifting and un lifting of weights onto a platform creates additional error. This error is due to the un controlled application of the weight onto the pan. Handling weights and placing them onto a pan may cause vibration error, tape misalignment error and introduce a pendulum or swinging effect. To minimize this error it is recommended to use a fixed chain loaded or pneumatic weight stack in which the weights can be applied as uniformly as possible. 93

94 Introduction of Bending and Parasitic effects The influence of bending and parasitic effects on a transducer will be dependent on the particular design of that transducer. However the reproducibility of the device, when calibrated in different orientations, can often be a good indicator of the influences of bending. 94

95 Calibrated beam deflection and temperature The material the beam is machined from will determine the appropriate amount of weight that can be supported by the beam. Using more weight than the beam is designed for may cause significant deflection. The properties of the material of the beam will expand and contract changing the length of the beam. The temperature effect can be minimized by maintaining the room temperature of the calibration laboratory at the same temperature the beam was calibrated. Any temperature corrections can be made based on the temperature coefficient of the material the beam. 95

96 Misalignment Alignment - practically it is very difficult to perfectly align a transducer in a torque machine. Usually there will be some misalignment due to the mismatch of the two axes that will give a radial, angular, or axial misalignment or any combination of the three. One way to minimize this is through the use of flexible coupling elements. 96

97 Misalignment (Cont.) What are the actual effects of a misalignment error? What does this do to the combined uncertainty? 97

98 Combined Type A and B Uncertainty Example assuming 1% misalignment error. Type A and B uncertainty analysis COMPANY TORQUE CELL MANUFACTURER TORQUE CELL S/N CAPACITY ASTM E2428 Uncertainty for K=2 THE LOWEST FORCE AT WHICH THE SECONDARY STANDARD WILL BE USED PRIMARY TORQUECALIBRATION STANDARD UNCERTAINTY K=1 PRIMARY ELECTRICAL CALIBRATION STANDARD UNCERTAINTY K=1 (IF APPLICABLE) CAL DATE CALCULATED VALUES TORQUE CELL UNCERTAINTY IN % FOR FULL SCALE K= % TORQUE CELL UNCERTAINTY IN % FOR FULL SCALE K= % TORQUE CELL UNCERTAINTY IN % FOR LOWEST FORCE APPLIED K = % TORQUE CELL UNCERTAINTY IN % FOR LOWEST FORCE APPLIED K = % MOREHOUSE MOREHOUSE SAMPLE 1000 N-m N-m 100 N-m 0.010% 10/27/2010 Type A Uncertainty % Uncertainty Description Uncertainty Distribution Divisor Standard Uncertainty Squared ASTM E2428 Uncertainty % at the lowest calibration torque to be used % normal E E-08 Combined Type A Uncertainty 1.10E E-08 Type B Uncertainty % Uncertainty Description Uncertainty Distribution Divisor Standard Uncertainty Squared PRIMARY TORQUE CALIBRATION STANDARD UNCERTAINTY 0.010% rectangular E E-09 PRIMARY ELECTRICAL CALIBRATION STANDARD UNCERTAINTY (IF APPLICABLE) 0.000% rectangular E E+00 STABILITY OF THE SECONDARY TORQUE STANDARD OVER TIME 0.005% rectangular E E-10 MISALIGNMENT ERROR 1.000% rectangular E E-05 DISSEMINATION ERROR (FOR CALIBRATION LABORATORIES) 0.020% rectangular E E-08 TEMPERATURE ERROR +/- FROM CELL SPEC SHEET % rectangular E E-10 Combined Type B Uncertainty 5.78E E-05 FOR K= 1 UC= SQUARE ROOT OF TOTAL COMBINED TYPE A AND B % FOR K=2 U = K * UC (UNCERTAINTY % AT THE LOWEST FORCE TO BE APPLIED) 1.155% N-m 98 UC AT CAPACITY * 2 ( (UNCERTAINTY % AT INSTRUMENT CAPACITY) 1.155% N-m

99 Misalignment Error A misalignment error of.1% will more than triple the combined uncertainty. A misalignment error of 1% will raise the combined calibration uncertainty from.355 N-m at full scale to N-m. 99

100 Worn drives Transducer clamped in a mounting bench for an unsupported beam calibration. The fit of the squares and rigidity of the mounting plate are important factors. (note: the above picture is an example of a transducer properly mounted) If these fittings are worn the proper torque will not be generated. It is not uncommon for this error to exceed 10% of the desired torque value. 100

101 Bending Forces on an unsupported beam In this case it is important to be aware of the influence that bending can have on the measurement result and, where possible, to be able to quantify this. The mounting of the transducer is very important. The transducer should be held as rigidly as possible on the mounting plate. The fit of the squares should be as good as possible to minimize any slack 101

102 Bending Forces on an unsupported beam A Comparison Between Supported and Unsupported Beams For Use in Static Torque Calibrations by Andy Robinson and Barry Pratt revealed that the weight of unsupported calibration beams in relation to the capacity of the transducer is critical. Figure 8 shows: Transducer A (2kN-m transducer calibrated with 1.5 kn-m unsupported beam) Transducer B (200N-m transducer calibrated with.5 kn-m unsupported beam) Transducer A (1kN-m transducer calibrated with 1.5 kn-m unsupported beam) 102

103 Bending Forces on an unsupported beam Transducer B 200 N-m transducer was then calibrated with the 1.5kN arm. Figure 10: 500 N-m beam weighs 16.6 kg 1.5 kn-m beam weighs 23.4 kg 103

104 Bending Forces on an unsupported beam Conclusion: The weight of the unsupported calibration beam in relation to the capacity of the load cell has a significant effect on the transducer output. The worst affected region is 0 20% where the weight of the beam is most significant in relation to the torque being measured. Under certain conditions an unsupported beam achieved results within 0.02% of the reference value from the % range. (Transducer C which was a 2kN-m shaft type) 104

105 Bending Forces on an unsupported beam Bearings are often used to support the beam and eliminate bending effects due to the weight of the lever beam and weights. However depending on its quality the bearing itself may introduce errors such as friction and the performance of the bearing still requires evaluation 105

106 Verification by inter comparisons with a National Metrology Institute or Morehouse Verification of the torque system will allow the laboratory to diagnose and correct problems. Verification will give confidence in torque measurement. Verification can be done with equipment owned by Morehouse and rented to the end user or it can be done with an existing torque transducer using the Primary Torque Standard at Morehouse Instrument Company or at a NMI. 106

107 In Class Demonstration Primary torque standards a dead weight force applied through a lever arm or wheel, with a calibrated length or radius of a known uncertainty, that is traceable to national standards. 107

108 In Class Demonstration Morehouse Single Ended Beam Known Length of meters, Weight = N (We will tare this out for our example) Morehouse Force Calibrated Weights = LBF or N 108

109 In Class Demonstration We will be comparing the single end beam and mass calibration against a torque cell calibrated by our primary standard machine The uncertainty of the torque cell is.019 N-m At 40 N-m U =.0475% of applied At 80 N-m U =.0238% of applied At 120 N-m U=.0158% of applied 109

110 In Class Demonstration Expectations At 40 N-m we expect the cell to read m x = N-m within +/-.019 N-m At 80 N-m we expect the cell to read m x = N-m within +/-.019 N-m At 120 N-m we expect the cell to read m x = N-m within +/-.019 N-m Per ASTM E2428 our system should have an uncertainty better than.012% of applied when combining the uncertainty for force and length. In actuality the uncertainty on this system is.018%. We know length within +/-.005 in and weight within.003% (this includes gravity correction for this area) for K=2 We would expect the results to be accurate to +/-.019 N-m +/-.03% of applied for k=2. We must then reduce everything to K=1 and sum the squares 110

111 In Class Demonstration Expectations For us to have confidence in this system we should have agreement as follows for k=2 For K =1 Torque cell =.0095 N-m, Single Beam system =. 01% We must divide.0095 N-m by the applied torque and square it. Then we must add.01 ^ 2 and take the square root. Finally multiply by 2 and we should get the results below. At 40 N-m results should be better than.052% of applied = +/-.021 N-m At 80 N-m results should be better than.031% of applied = +/-.025 N-m At 120 N-m results should be better than.026% of applied = +/-.032 N-m 111

112 In Class Demonstration Expectations Target 40 N-m within +/-.021 N-m Actual = (this must be filled in from test) Target 80 N-m within +/-.025 N-m Actual = (this must be filled in from test) Target 120 N-m within +/-.032 N-m Actual = (this must be filled in from test) 112

113 In Class Demonstration Expectations Conclusion: Having not known the actual results from the test we are expecting that the uncertainty of this system is much greater than.02% of applied torque. Without verification of the system the end user would be performing torque calibrations while claiming their system uncertainty is better than.02% of applied and they would have the proper calibration certificates for length and force to back up this claim. Verification or proficiency testing against a torque standard from a NMI or lab that has verified their own equipment with a torque cell calibrated by a NMI should be done before any torque system is put into use and prior to any accreditation process. 113

114 What is so special about the NPL machine located at Morehouse Instrument Company? The next series of slides deal with the detail and work that went into our torque calibration laboratory The slides also include a detailed analysis of the NPL machine and the great lengths that were taken to achieve a torque uncertainty of 0.002% of applied torque 114

115 Morehouse Torque Calibration Laboratory 115

116 Construction of the Torque Calibration Laboratory. A 6 8 deep 12 x16 pit was dug inside our current building. The pit was reinforced and a center plinth constructed. The plinth features a heated floor to control the temperature. NPL did not have heated flooring. Since we knew some of the problems NPL had with airflow we designed our room to keep any air from blowing directly on the machine. From preliminary testing we have been able to maintain temperature at 20 degrees Celsius +/- 0.2 Degrees Celsius. 116

117 Installation of NPL s Primary Torque Standard The Machine arrived at Morehouse Mannie Panesar and Michael Harrison from NPL arrived to start the installation. Andy Robinson and Andy Knott (pictured bottom right) from NPL arrived to verify the machine was within 0.002% of applied torque. 117

118 Target Uncertainty NPL s Torque machine target specification Torque range Deadweight loading 2 N m 2 kn m Target uncertainty 1 x 10-5 Clockwise and anti-clockwise Incremental and decremental Working envelope 1 m 3 118

119 Conceptual design 2 kn m lever deadweight machine Realised uncertainty % Vertical design pure torque generated via identical weightstacks located at either end of the lever beam Twin beam carbon fibre lever arm mounted on a central air bearing 119

120 Conceptual design Vertical torque axis necessitates a change in force direction Vertical force generated by the weightstacks is converted to a horizontal force connecting to the lever arm Achieved using boron fibre tapes and a pulley air bearing 120

121 Conceptual design Lever beam can be rotated independently of the torque transducer Pulley system is castored to allow rotation about its vertical axis The same beam and weightstacks used for clockwise and anticlockwise measurements 121

122 Conceptual design Ranges Primary: 20 N m to 2 kn m in 30 steps Secondary: 1 N m to 100 N m in 28 steps Other features Continuous torque calibration Ability to introduce and measure misalignments Asymmetric torque mode 122

123 Air bearings Criteria Requirement to provide both radial and axial stiffness to support the predicted forces. Minimal friction target (0.1 to 0.5) µn m (Friction is a combination of boundary and viscous shear of air film plus turbine torque) 123

124 Air bearings Central pivot air bearing Double thrust and air bearing system Upper bearing enables rotation of the lever beam Lower bearing supports the transducer Incorporates a drag cup damper to compensate for the natural frequency of the system 124

125 Air bearings Boundary and viscous shear influences are time dependant and so can be ignored Turbine torque measured with calibrated micro air jet 220 µn m This is constant and is tared so target uncertainty is not exceeded Castor pulley bearing system provides horizontal and vertical rotation Designed to carry a load of 1 kn transmitted by suspension tape at supply pressure of 345 kpa (50 psi) 125

126 The lever and reaction beam Designed by NPL materials team Criteria Lightweight construction to reduce inertia Very low deflection under load Thermal expansion of beam considered critical 126

127 The lever and reaction beam Two parallel 2 m long highmodulus carbon fibre tubes Stainless steel end fittings and central boss Calculated coefficient of thermal expansion = 0.3 x 10-6 K -1 (Steel is 16 x 10-6 K -1 m/m) Finite element analysis used to provide confidence in design 127

128 The lever and reaction beam Beam has nominal m length measured to a relative uncertainty of 3.3 x 10-6 Length is monitored using a comparator With an uncertainty of 2.0 x 10-6 data show that length has changed by less than 3 µm over a 1 m length Overall length uncertainty contribution x

129 Boron fibre tapes Criteria No deformation under load Lightweight Flexible 129

130 Boron fibre tapes Epoxy pre-impregnated tape of 150 µm boron fibre tapes Chemical vapour deposition of a boron fibre gas on to fine tungsten wire Sandwiched and bonded between a titanium end connector 5 mm tape used for torques up to 100 N m 20 mm tape used for torques up to 2 kn m 130

131 Boron fibre tapes Tape is required to change orientation 20 mm tape has 90 joint made from carbon and glass fibre 5 mm tape has a pre-formed twist 131

132 Control Software Graphic user interface Visual Basic 6.0 Several working modes Displays real-time machine outputs Complex motion profiles weightstack platforms slow as each weight is added or removed The lever beam is re-datumed after each weight increment Built In safety features 132

133 Couplings Adaptability - capable of incorporating most shaft, flange, and square drive transducers Based around a hydraulic friction joint Incorporates Rexnord flexible couplings eliminating radial, axial, and angular misalignments 133

134 Performance & Uncertainty Relative expanded uncertainty 2 x 10-5 Traceable to base SI units of mass, length, and time Major contributors Force & alignment of force Length Alignment of boron tapes Validated by NPL through intercomparison work with other labs Validated by NPL at Morehouse through intercomparison with work done by NPL before the machine was dismantled and then after the machine was reassembled at Morehouse. 134

135 Installation of NPL s Primary Torque Standard Cal Factor / mv/v per N.m Raute 2kNm ,200 1,600 2,000 Torque / N.m NPL MIC Difference Incremental Difference / % Comparison graph showing agreement within % of applied torque with data taken before decommissioning the machine at NPL and after installation at Morehouse 135

136 Conclusion Commissioning completed in December 2005 at NPL Commissioned May 2010 at Morehouse Instrument Company Achieved target of realizing uncertainties at world leading levels via an novel and innovative design Dissemination of the unit of torque expected to have a significant impact in US industry Flexibility of the machine will provide a powerful research tool 136

137 This concludes this section on torque calibration Any Questions? 137