40
1 Optimizing Strain Gage Optimizing Strain Gage Load Cell Load Cell Performance Performance by LaVar Clegg by LaVar Clegg Interface, Inc. Interface, Inc. Western Regional Strain Gage Committee Western Regional Strain Gage Committee Conference September 13, 2011 Conference September 13, 2011

Optimizing Load Cell Performance

Embed Size (px)

DESCRIPTION

This paper explores the fundamental characteristics of medium to high capacity load cells and how they are affected by the types and implementation of strain gages.

Citation preview

Page 1: Optimizing Load Cell Performance

1

Optimizing Strain Gage Optimizing Strain Gage Load CellLoad Cell

PerformancePerformance

by LaVar Cleggby LaVar CleggInterface, Inc.Interface, Inc.

Western Regional Strain Gage Committee Conference Western Regional Strain Gage Committee Conference September 13, 2011September 13, 2011

Page 2: Optimizing Load Cell Performance

2

SummarySummary

This paper explores the fundamental This paper explores the fundamental characteristics of medium to high capacity characteristics of medium to high capacity load cells and how they are affected by the load cells and how they are affected by the types and implementation of strain gagestypes and implementation of strain gages

Guidance for successful application of Guidance for successful application of load cellsload cells

Page 3: Optimizing Load Cell Performance

3

Load Cell Types RepresentedLoad Cell Types Represented

Low Profile, center thread Low Profile, flange mount

Single Column Hollow column

Page 4: Optimizing Load Cell Performance

4

Performance FeaturesPerformance Features

1. Linearity over measuring range1. Linearity over measuring range

2. Hysteresis (descending loads)2. Hysteresis (descending loads)

3. SEB (why is it useful ?)3. SEB (why is it useful ?)

4. Output symmetry4. Output symmetry

5. Rejection of extraneous loads5. Rejection of extraneous loads

6. Sensitivity to mounting6. Sensitivity to mounting

Page 5: Optimizing Load Cell Performance

5

1. Linearity1. Linearity

A.A. Important because load cell non-Important because load cell non-linearity represents system error when the linearity represents system error when the instrumentation is linear, as it typically isinstrumentation is linear, as it typically is

B.B. Only smooth calibration curves can Only smooth calibration curves can be be corrected by compensation in the corrected by compensation in the instrumentationinstrumentation

Page 6: Optimizing Load Cell Performance

6

Low Profile Linearity ExampleLow Profile Linearity Example

We use a 10,000 lbf low profile type load We use a 10,000 lbf low profile type load cell to examine excellent linearity cell to examine excellent linearity behavior over a wide measuring rangebehavior over a wide measuring range

Page 7: Optimizing Load Cell Performance

7

Low Profile ConstructionLow Profile Construction

Shear beam gage shown Shear beam gage shown

Load Surface

Shear beam

Gage

Base

Page 8: Optimizing Load Cell Performance

8

Compression CalibrationCompression Calibration

Page 9: Optimizing Load Cell Performance

9

10K Compression (cont’d)10K Compression (cont’d)

a. Nonlinearity is relatively low 0.02%FSa. Nonlinearity is relatively low 0.02%FS

b. Calibration points fit a polynomial curve b. Calibration points fit a polynomial curve very closelyvery closely

c. Curve is smooth clear down to zero load c. Curve is smooth clear down to zero load (tested over 10 to 10,000 lbf range)(tested over 10 to 10,000 lbf range)

Page 10: Optimizing Load Cell Performance

10

Expanded Scale (10X magnified)Expanded Scale (10X magnified)

Page 11: Optimizing Load Cell Performance

11

Comparable behavior in tensionComparable behavior in tension

Page 12: Optimizing Load Cell Performance

12

Expanded Scale (10X magnified)Expanded Scale (10X magnified)

Page 13: Optimizing Load Cell Performance

13

Column Cell LinearityColumn Cell Linearity

a.a. Nonlinearity relatively Nonlinearity relatively large due to the expansion large due to the expansion or contraction of the or contraction of the column column diameter with loaddiameter with load

b.b. But well-behaved But well-behaved smooth smooth calibration curves calibration curves normally normally fitting a 2fitting a 2ndnd degree degree

polynomialpolynomial c.c. Tension and Tension and

compression compression opposite opposite polarity of non-polarity of non- linearitylinearity

Page 14: Optimizing Load Cell Performance

14

2160-1000 kN (225 Klbf) Example2160-1000 kN (225 Klbf) Example

Tension NL = - 0.073%FS

Compression NL = +0.053%FS

Page 15: Optimizing Load Cell Performance

15

Strain Gage InfluenceStrain Gage Influence

The preceding example load cell was The preceding example load cell was made with modified-Karma alloy gagesmade with modified-Karma alloy gages

Constantan alloy gages tend to produce Constantan alloy gages tend to produce higher nonlinearity, about 0.10%FS, in our higher nonlinearity, about 0.10%FS, in our experienceexperience

However, a generalization should not be However, a generalization should not be made without considering all effects of made without considering all effects of geometry and transverse gage factor geometry and transverse gage factor

Page 16: Optimizing Load Cell Performance

16

Flange Mount Low Profile LinearityFlange Mount Low Profile Linearity

a.a. Nonlinearity relatively low, benefiting from the Nonlinearity relatively low, benefiting from the solid solid hubhub

b.b. But well-behaved smooth calibration curves, fit But well-behaved smooth calibration curves, fit 22ndnd or 3or 3rdrd degree polynomial degree polynomial

c.c. Tension – compression symmetry is excellentTension – compression symmetry is excellent

Page 17: Optimizing Load Cell Performance

17

1238 - 250 kN Flange Mount Example1238 - 250 kN Flange Mount Example

Tension NL = - 0.008%FS

Compression NL = - 0.010%FS

Page 18: Optimizing Load Cell Performance

18

Hollow Column LinearityHollow Column Linearity

a.a. Nonlinearity slightly better than single columnNonlinearity slightly better than single column b.b. But well-behaved smooth calibration curves, fit But well-behaved smooth calibration curves, fit

33rdrd degree polynomialdegree polynomial c.c. Tension – compression symmetry is goodTension – compression symmetry is good

Page 19: Optimizing Load Cell Performance

19

2350 - 2000 kN Hollow Column Example2350 - 2000 kN Hollow Column Example

Tension NL = - 0.070%FS

Tension NL = - 0.032%FS

Page 20: Optimizing Load Cell Performance

20

2. Hysteresis2. Hysteresis

A.A. Often misunderstoodOften misunderstood B.B. Descending calibration points are Descending calibration points are

valid valid only for the particular FS value only for the particular FS value of a testof a test

C.C. Nevertheless, the measure of Nevertheless, the measure of hysteresis has value as an indicator of hysteresis has value as an indicator of the range of error to expect from load the range of error to expect from load points that do not necessarily ascend points that do not necessarily ascend from zerofrom zero

Page 21: Optimizing Load Cell Performance

21

Example of good hysteresis behaviorExample of good hysteresis behavior

Same 10,000 lbf low profile type load cell Same 10,000 lbf low profile type load cell we examined for linearitywe examined for linearity

Descending curve as well behaved as Descending curve as well behaved as the ascending curvethe ascending curve

Page 22: Optimizing Load Cell Performance

22

Smooth descending curveSmooth descending curve

H = +0.03%FS and descending curve closes at H = +0.03%FS and descending curve closes at zero load. Closure requires well behaved zero load. Closure requires well behaved hysteresis and very low creephysteresis and very low creep

Page 23: Optimizing Load Cell Performance

23

Expanded Scale (10X magnified)Expanded Scale (10X magnified)

Page 24: Optimizing Load Cell Performance

24

Many levels of performanceMany levels of performance

In calibrating load cells from many In calibrating load cells from many manufacturers around the world, it is seen manufacturers around the world, it is seen that not all are as well-behaved as the that not all are as well-behaved as the preceding examples of Interface cellspreceding examples of Interface cells

The differences are in the subtleties of The differences are in the subtleties of design and quality control design and quality control

Page 25: Optimizing Load Cell Performance

25interface

Example of lower quality load cellExample of lower quality load cell(not an Interface load cell)(not an Interface load cell)

A. Nonlinear near zero load

B. High hysteresis

C.Non-closure of zero return indicates high creep

D.Takes a 4th degree polynomial to fit a curve

Page 26: Optimizing Load Cell Performance

26

3. SEB3. SEB

Static Error Band (SEB) is often misunderstood. The SEB Static Error Band (SEB) is often misunderstood. The SEB output line provides a single slope calibration constant that output line provides a single slope calibration constant that minimizes error on average over a force range.minimizes error on average over a force range.

Demonstration of SEB vs Terminal Output

-0.060

-0.050

-0.040

-0.030

-0.020

-0.010

0.000

0.010

0.020

0 20 40 60 80 100

Load (%FS)

Err

or

fro

m s

tra

igh

t lin

e (

%F

S)

Data points Terminal Output Line SEB Output Line

Ascending

Descending

Page 27: Optimizing Load Cell Performance

27

4. Output Symmetry4. Output Symmetry

Important when both tension and compression Important when both tension and compression loadings use the same instrumentation gainloadings use the same instrumentation gain

Generally, low profile shear cells better than Generally, low profile shear cells better than column cellscolumn cells

Symmetry Error of our example cells:Symmetry Error of our example cells:

Low Profile 10 Klbf 0.01%Low Profile 10 Klbf 0.01%

LP Flange 200 kN 0.03%LP Flange 200 kN 0.03%

Hollow column 2000 kN 0.05%Hollow column 2000 kN 0.05%

Single Column 1000 kN 0.25%Single Column 1000 kN 0.25%

Page 28: Optimizing Load Cell Performance

28

5. Rejection of extraneous loads5. Rejection of extraneous loads

It is desired to measure FzIt is desired to measure Fz Fx, Fy, Mx, My, Mz are extraneousFx, Fy, Mx, My, Mz are extraneous

Fz

Fx

Fy

Mz

Mx

My

Page 29: Optimizing Load Cell Performance

29

Axial Load vs. Eccentric LoadsAxial Load vs. Eccentric Loads

Axial EccentricAxial Eccentric

Page 30: Optimizing Load Cell Performance

30

Method of testing eccentric load sensitivityMethod of testing eccentric load sensitivity

Sensitivity of less Sensitivity of less than 0.1% / inch than 0.1% / inch is achieved on is achieved on shear low profile shear low profile type cellstype cells

A force is applied on a moment arm while monitoring load cell output

Page 31: Optimizing Load Cell Performance

31

Eccentric adjustment exampleEccentric adjustment example

1020-25K Eccentric Load Plot

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0 30 60 90 120 150 180 210 240 270 300 330 360

Position

Err

or

(% /

inch

)

Initial Adjusted

Page 32: Optimizing Load Cell Performance

32

6. Sensitivity to mounting6. Sensitivity to mounting

Degrees of reducing installation influenceDegrees of reducing installation influence

Basic cell Factory-installed base

Factory-installed stud

Integral machined stud

D e c r e a s i n g m o u n t i n g s e n s i t i v i t y

Page 33: Optimizing Load Cell Performance

33

Preloading large threadsPreloading large threads

Challenging, as in this 10MN wire rope testChallenging, as in this 10MN wire rope test

10 MN (2.2 Million Pound) load cell

Page 34: Optimizing Load Cell Performance

34

Advantage of flange load cellsAdvantage of flange load cells

Screws can be installed with conventional Screws can be installed with conventional torque wrenches or hydraulic torque torque wrenches or hydraulic torque wrencheswrenches

Page 35: Optimizing Load Cell Performance

35

Example of mounting sensitive load cellExample of mounting sensitive load cell(not an Interface cell)(not an Interface cell)

A. Nonlinear near zero load

B. High hysteresis

C.Requires hard bearing plates

D.No polynomial fit

100 Klbf

Page 36: Optimizing Load Cell Performance

36

Finite Element Analysis of MountingFinite Element Analysis of Mounting

Performance is always dependent upon fixation of the load cell to its Performance is always dependent upon fixation of the load cell to its live end and dead end structures. Here in this cutaway view the live end and dead end structures. Here in this cutaway view the screw clamping force is being analyzed.screw clamping force is being analyzed.

Page 37: Optimizing Load Cell Performance

37

Strain Gage AlignmentStrain Gage Alignment

Alignment of gages is critical to rejection of Alignment of gages is critical to rejection of extraneous loads. Here the gage can be seen extraneous loads. Here the gage can be seen well-aligned with the precisely applied scribe linewell-aligned with the precisely applied scribe line

Page 38: Optimizing Load Cell Performance

38

Tension LinkTension Link

Another type of high capacity load cellAnother type of high capacity load cell Load is measured between clevis pins in Load is measured between clevis pins in

the ends of the cellthe ends of the cell

Page 39: Optimizing Load Cell Performance

39

Tension Link GagingTension Link Gaging

Strain gages are strategically placed to Strain gages are strategically placed to sense tensile stress in the web sectionsense tensile stress in the web section

Page 40: Optimizing Load Cell Performance

40

ConclusionConclusion

Medium to high capacity load cells are Medium to high capacity load cells are successfully made in a variety of stylessuccessfully made in a variety of styles

The appropriate style is determined by the The appropriate style is determined by the demands of the applicationdemands of the application

Measurement performance and Measurement performance and environmental needs can be met through environmental needs can be met through good engineering design and good engineering design and manufacturing of load cells manufacturing of load cells