National Aeronautics andSpace Administration

NASA Technical Memorandum 104331

A Limited In-Flight Evaluation of the Constant Current Loop Strain Measurement Method

Candida D. Olney and Joseph V. Collura

August 1997

National Aeronautics andSpace Administration

Dryden Flight Research CenterEdwards, California 93523-0273

1997

NASA Technical Memorandum 104331

A Limited In-Flight Evaluation of the Constant Current Loop Strain Measurement Method

Candida D. Olney and Joseph V. Collura

NASA Dryden Flight Research CenterEdwards, California

A LIMITED IN-FLIGHT EVALUATION OF THE CONSTANT CURRENT LOOP STRAIN MEASUREMENT METHOD

Candida D. Olney* and Joseph V. Collura†

NASA Dryden Flight Research CenterEdwards, California

Abstract

For many years, the Wheatstone bridgehas been used successfully to measureelectrical resistance and changes in thatresistance. However, the inherent problemof varying lead wire resistance can causeerrors when the Wheatstone bridge is usedto measure strain in a flight environment.The constant current loop signal-conditioning card was developed toovercome that difficulty. This paperdescribes a limited evaluation of theconstant current loop strain measurementmethod as used in the F-16XL ship 2Supersonic Laminar Flow Control flightproject. Several identical strain gages wereinstalled in close proximity on a shockfence which was mounted under the leftwing of the F-16XL ship 2. Two strain gagebridges were configured using theconstant current loop, and two wereconfigured using the Wheatstone bridgecircuitry. Flight data comparing the outputfrom the constant current loop configuredgages to that of the Wheatstone bridgeswith respect to signal output, error, andnoise are given. Results indicate that theconstant current loop strain measurementmethod enables an increased output,unaffected by lead wire resistancevariations, to be obtained from straingages.

* Aerospace Engineer, Aerostructures Branch, 805-258-3988†Electrical Engineer, Instrumentation Branch, 805-258-2814

This paper is declared a work of the U.S. Government and is n

Nomenclature

A gain

E excitation voltage across the bending bridge, volts

maximum expected voltage output of the gage, volts

GF gage factor

I current, ampere

PCM pulse code modulator

resistance of strain gage… nth, ohms

change in resistance in strain gage … nth, ohms

resistance of the reference strain gage, ohms

resistance of lead wire … nth, ohms

SLFC Supersonic Laminar Flow Control

voltage drop across inboard strain gages, volts

voltage drop across outboard strain gages, volts

voltage drop across strain gage … nth, volts

voltage output, volts

voltage output of the strain gage … nth, volts

E omax

Rg

∆Rg

Rref

Rw

V a

V b

V g

V out

V out g

ot subject to copyright protection in the United States.

maximum allowable voltage input to the PCM, volts

Vrms root mean square voltage

β angle of sideslip, deg

ε strain, µin/in

expected maximum strain, µin/in

µin/in microstrain, 10–6in/in

Introduction

For many years, strain gage instrumenta-tion has been used to measure strainwithin structural components. Traditionally,these strain gages were configured usingthe Wheatstone bridge circuit. However,the output of Wheatstone bridges issusceptible to varying lead wire resistancebecause of temperature effects, lead wiregauge, and wire length. These variationsmay decrease the sensitivity of the straingage to strain inputs (ref. 1).

A constant current loop strain measure-ment method was developed by NASADryden Flight Research Center (DFRC),Edwards, California, and patented by KarlF. Anderson (ref. 2) to overcome thedifficulties associated with the Wheatstonebridge. This constant current loop strainmeasurement method consists of the straingages, constant current loop signal-conditioning card, pulse code modulator(PCM), and signal transmitter. The constantcurrent loop signal-conditioning card uses aconstant current source in conjunction withstrain gages to measure strain.

This signal-conditioning card was initiallytested in the Thermostructural ResearchFacility (ref. 3).* During these laboratorytests before flight, Wheatstone bridgesignal conditioning was replaced with theconstant current loop signal conditioning.

Results indicated that the constant currentloop strain measurements were insensitiveto lead wire resistance changes and wereaccurate when connected to a strainindicator calibration standard. However,data comparisons between the Wheatstonebridge and the constant current loop strainmeasurement systems were not made.

Flight tests were performed to compare thestrain gage outputs using the constantcurrent loop and Wheatstone bridgestrain measurement methods in a flightenvironment. The constant current loopstrain measurement system was flighttested using three different lead wireconfigurations: seven-, five-, and three-wireconfigurations. Diagrams of these threeconfigurations are presented in appendixA. The seven-wire configuration allowsindividual gage measurements to be made.All three configurations were used to obtainstrain outputs from four-active-arm-bendingbridges.

This paper presents a comparison of theWheatstone bridge and constant currentloop signal-conditioning methods and adescription of the experimental method. Inaddition, a comparison of the data collectedduring flights 87 and 88 of the F-16XLship 2 Supersonic Laminar Flow Control(SLFC) project using the Wheatstonebridge and the constant current loop strainmeasurement methods is given. Use oftrade names or names of manufacturers inthis document does not constitute anofficial endorsement of such products ormanufacturers, either expressed or implied,by the National Aeronautics and SpaceAdministration.

*The name of the DFRC Thermostructural ResearchFacility was changed to the Flight Loads Laboratoryin 1996.

V PCMmax

εmax

2

Aircraft Configuration

The constant current loop signal-conditioning card was installed as asecondary experiment on the F-16XLship 2 SLFC project (ref. 4) (fig. 1). TheSLFC experiment was designed to producelaminar flow over a gloved test section onthe left wing of the modified F-16XL ship 2by using active suction through the gloveto remove the boundary layer duringsupersonic flight. It was determined thatduring supersonic flight, a shock wave fromthe engine inlet distorts the flow on theleading edge of the wing. A shock fence(fig. 2) was installed under the wing toprevent the shock wave from interferingwith the airflow over the wing.

Figure 3 shows an overall view of theinboard surface of the shock fence and therelative positions of the strain gages. Twosets of prime and spare four-active-arm-bending bridges were installed on the

shock fence in the areas of the maximumpredicted strain. The prime and spare straingages were located in close proximity toeach other and originally were wired in aWheatstone bridge configuration. One setwas located at a forward location on theshock fence, and one was installed at an aftlocation. To perform a flight evaluation, thesignal-conditioning cards for the spareWheatstone bridges were replaced withconstant current loop cards.

The spare strain gages at the forwardlocation were re-wired in the constantcurrent loop seven-lead-wire configura-tion. The output from the seven-wireconfiguration was used to create output forthe five- and three-wire configurations. Thewiring for the spare strain gages at the aftlocation were configured using fivelead wires. These configurations allowedcomparisons of the output of the two strainmeasurement systems to be made.

3

EC95 43297-07Figure 1. The F-16XL ship 2.

Figure 2. The F-16XL ship 2 shock fence.

Figure 3. Inboard surface of the shock fence with strain gages installed.

Aft bridge location

Forward bridge location

970659

Gage 1

Strain gages used in constant current loop circuit

Not to scale

Gage 2

Forward

970660

Strain gages used in Wheatstone bridge circuit

Thermocouple

4

Description of Wheatstone Bridge and Constant Current Loop Signal Paths

Figure 4 shows the signal path for all ofthe strain measurement systems. Afterexcitation is applied to the strain gages, thesignal is carried to the signal-conditioningcards and processed. The signal then issent to the PCM and is downlinked tothe ground station. Several significantdifferences exist between the signal paths

of the constant current loop and theWheatstone bridge strain measurementsystems. These differences include thegage excitation method, number oflead wires required, signal-conditioningprocess, and wiring to the PCM.

Excitation Methods

The excitation methods are fundamentallydifferent for the two measurement systems.

Figure 4. Instrumentation signal path for strain gage measurements.

Signal-conditioning

cards

Filtered andamplified

gage signalExcitation

Gagesignal

970661

Straingages

Pulsecode

modulator

Transmitterdownlink

Groundstation

The Wheatstone bridge method uses aconstant voltage source for the excitation ofthe strain gages. The Wheatstone bridgesignal-conditioning applies 10 volts acrossthe bridge, which results in 5 volts acrossthe individual gages. A diagram of theWheatstone bridge circuit used in thisexperiment is presented in appendix A. Asthe lead wire resistances increase, thevoltage potential across the gages drops,thereby causing a reduced effectiveexcitation voltage. Therefore, the sensitivityof the Wheatstone bridge to strain changesdecreases when the lead wire resistancesincrease.

As the name implies, the constant currentloop signal-conditioning card uses constantcurrent regulation to provide gageexcitation. The regulator compares a stable5-volt source with the voltage across thereference resistor in its current loop (ref. 3).Because the reference resistor is physicallylocated on the signal-conditioning card,it does not experience the sameenvironmental conditions that the straingages experience. Therefore, theresistance of the reference resistor isconstant. The return current input into thecircuit is adjusted by the regulator so that astable, constant current is applied to thestrain gages.

Lead Wires

Another difference between the constantcurrent loop and Wheatstone bridgeconfigurations is the number of lead wiresrequired. The two Wheatstone bridgeseach had four lead wires contained inshielded, four-wire bundles. This approachcauses the lead wires to experienceessentially the same environment so thattheir resistances change consistently. If thelead wires do not change the same, thenthe strain measurements will not beaccurate (ref. 5). However, even if the leadwire resistances do increase equally, thesensitivity of the strain gages to a straininput is reduced by a varying amount whichis difficult to quantify.

The constant current loop strain gageswere arranged with three different lead wireconfigurations for this experiment. Thesethree configurations use seven, five, andthree lead wires. In this application,the forward constant current loop straingages were wired in the seven-lead-wireconfiguration. The five- and three-lead-wireconfigurations were created on the signal-conditioning card by jumpering the signalsat the appropriate points. Appendix A givesa derivation of the outputs of the constantcurrent loop strain gages and explains how

5

data from the five- and three-lead-wireconfigurations were obtained.

The seven lead wires were included ina shielded, eight-wire bundle, therebyleaving one wire unused. The aft constantcurrent loop gages were configured usingfive lead wires contained in a shielded, six-wire bundle. The derivation in appendix Areveals that no assumptions need to bemade about the lead wire resistancechanges in the seven- and five-lead-wireconfigurations. However, when the three-lead-wire configuration is used, theassumption must be made that the twocurrent-carrying lead wires will changeequally. As with the Wheatstone bridgeconfiguration, this assumption is valid aslong as the current-carrying lead wiresexperience the same environment.However, unlike the Wheatstone bridge, achange in the lead wire resistances doesnot affect the output of the constant currentloop.

Signal-Conditioning Cards

The outputs of the strain gages are sent totheir respective signal-conditioning cards.The signal-conditioning cards for the twostrain measurement systems differ inthe number of signal-conditioning cardsrequired, number of measurements thatthese cards can accommodate, shuntresistor calibration capability, and signal-conditioning components.

The Wheatstone bridge signal conditioningin this evaluation used two cards (fig. 5(a)).Each set of two cards accommodates eightgage arrangements or bridges. The firstcard in the Wheatstone bridge signal pathis the shunt resistor calibration card. Thiscalibration card receives the voltageinformation from the full-bridge gageoutput. The full-bridge output may be

commanded to a simulated strain level byshunting a known resistance across onearm of the bridge. The resistor calibrationcard also allows for an external switch tocontrol all of the relays. These relayssimultaneously place all circuits intocalibration mode.

The second Wheatstone bridge cardamplifies the input signal and filters theoutput signal. The differential voltageacross the output of the bridge is readusing an instrumentation amplifier whichapplies the assigned gain to the signal. Theoutput from the amplifier is then sent to thefilter circuit. The filter is a third-order, active,low-pass filter which accomplishes noisereduction and presample filtering. Theoutput of the low-pass filter is then sent tothe PCM.

The bridges using the constant currentloop have the signal path shown in figure5(b). They use a single, multilayer, signal-conditioning card which can read fourfull-bridge strain gage measurements.These four measurements could includefour of the five- or three-lead-wire bendingbridges or eight individual gage measure-ments, which would be used in two of theseven-lead-wire bending bridges.

Unlike the Wheatstone bridge, the constantcurrent loop strain measurement systemused in this experiment did not provide theability to externally place the strain gagesinto calibration mode. A shunt calibrationswitch was located directly on the signal-conditioning card. Therefore, aircraft panelshad to be removed to perform the shuntcalibration. This time consuming processcan invalidate the aircraft preflight andprevent calibrations from being performedin flight.

6

(a) Wheatstone bridge.

(b) Constant current loop.

Figure 5. Signal-conditioning card signal paths.

Subtractoroperational

amplifier

ToPCM

Fromgages

Signal-conditioning

card

970662

Resistorcalibrationswitches

Resistorcalibration

card

Third-orderactive

low-pass filter

Subtractoroperationamplifier

ToPCM

Fromgages

970663

First-orderpassive

low-pass filter

First-orderpassive

low-pass filter

Dataoperationamplifier

The constant current loop signal-conditioning card used provided the abilityto wire the signal-conditioning card directlyto a shunt calibration card. Howeverbecause of the number of cards required forthe constant current loop measurements,no space was available for the additionalcalibration card. For example, to externallyput eight full bridges into calibrationmode, the constant current loop methodrequires three cards; meanwhile, theWheatstone bridge requires two. Tomeasure the voltage output from thesestrain gages, the constant current loopsignal-conditioning card was equipped witheight input instrumentation amplifiers foruse as subtractors and eight outputinstrumentation amplifiers to provide thegain.

Figure 5(b) shows the relationship of filtersand amplifiers for the constant current loop,and appendix A shows the placement ofthe amplifiers within the circuit. The outputsignal from the strain gage is routedthrough a first-order, passive, low-passfilter to the input of the subtractorinstrumentation amplifier. When using the

five- and three-lead-wire configurations,two instrumentation amplifiers are neededto take readings across both sides of thebridge.

However, in the seven-lead-wire configura-tion, one subtractor is required for readingthe voltage drop across each gage. Outputof the subtractor stage is then sent toanother instrumentation amplifier whoseinput is filtered with another first-order,passive, low-pass filter. This instrumenta-tion amplifier, or data amplifier, is responsi-ble for amplifying its input signal from eitherone or two subtractors to the level requiredby the PCM system input. The constantcurrent loop signal-conditioning card doesnot use a filter after the data amplifiers.

Wiring from Signal-Conditioning Card to PCM

Wiring schemes from the Wheatstonebridge and constant current loop signal-conditioning cards to the PCM differ in thenumber of data amplifier reference linesthat are tied together at each of the signal-conditioning cards. One PCM analog

7

module has 16 single-ended inputsconsisting of one return line for four inputlines.

For each Wheatstone bridge measure-ment, four signal lines are sent to the PCM,and four signal returns are jumperedtogether in one return line to the PCM(fig. 6(a)). Conversely, for the constantcurrent loop signal conditioning cards, alleight outputs of the card have their outputreturns tied together on the card. Only oneline is used to carry the return voltage(fig. 6(b)).

Experimental Method

Sixteen CEA-13-250-UW-350 (Micro-Measurements, Romulus, Michigan) straingages with a gage factor of 2.115 wereused in four-active-arm-bending bridges.Two complete bending bridges each wereplaced in close proximity in forward and aftlocations on the shock fence (fig. 3). Eachbridge consisted of two active arms on theinboard surface of the shock fence and twoon the outboard surface.

Figure 7 shows the forward gages installedon the inboard side of the shock fence. The

8

(a) Wheatstone bridge.

(b) Constant current loop.

Figure 6. Wiring from the signal-conditioning card to the pulse code modulator.

970664

Wheatstonebridge signal-conditioning

card

Pulse codemodulator

Vout1

Vout4Return

Vout5

Vout8Return

Input 1

Input 4

Return

Input 5

Input 8

Return

970665

Constantcurrent loop

signal- conditioning

card

Pulse codemodulator

Vout1

Vout8Return

Input 1

Input 8

Return

strain gages using the Wheatstone bridgecircuitry are forward of the strain gagesusing the constant current loop in both theforward and aft locations on the shockfence. A thermocouple was installed aft ofthe constant current loop configured straingages in the forward location to measurethe temperature of the shock fence duringthe test flights. Data from the thermocoupleare presented in appendix B.

The forward constant current loop bridgewas wired using the seven-lead-wireconfiguration. However, during the flight,five- and three-lead-wire measurementswere made concurrently at the forwardlocation. These simultaneous measure-ments were made by jumpering the signalsat the appropriate locations on theconstant current loop signal-conditioningcard (appendix A).

To properly scale the outputs from theconstant current loop strain gages, the gainfor the data amplifier was based upon themaximum strain level seen in previousflights. The maximum strain output was262 µin/in. Therefore, the strain scale forthe gage output was set to ±300 µin/in tomaximize the data resolution whilepreventing the signals from going off scale.Equation (1) gives the method used todetermine the maximum output for a singlegage.

(1)

Because E is set equal to the maximuminput of the PCM, 5 V, the gage factor is2.115 and is equal to 300 µin/in, themaximum output from one strain gage,

, is 3 mV. The next step is tocorrelate the maximum expected outputfrom the strain gage with the maximumallowable input for the PCM. This

E omax GF εmax E××=

εmax

E omax

9

Figure 7. The strain gages and thermocouple installed on the inboard surface of the shockfence at the forward location.

Gage 1

Thermocouple

Strain gages used inconstant current loop circuit

Strain gages used inWheatstone bridge circuit

Gage 2

Forward

970666

correlation factor is called the gain. Thegain, A, is calculated by

(2)

Because the input range for the PCM is±5 V, the gain for the individual gages iscalculated to be 1667. The output from thefour-gage bridge circuit is four times higherthan that of the individual gages. Therefore,the gains for the complete bridges arereduced by a factor of 4, and the gain forthe bending bridge measurements is 417.The gain on the Wheatstone bridge signal-conditioning card was set equal to the gainon the constant current loop signal-conditioning card to compare the signal-to-noise ratios.

Strain measurements were made at theforward and aft locations throughout theentire flight (fig. 3). However, because theshock fence was expected to experiencethe maximum strain when the angle ofsideslip, β, reached its maximum value, itwas decided that the strain measurementswould be compared using data taken whilethe aircraft completed three 5° nose-leftand nose-right β sweeps.

This maneuver was performed shortly aftertakeoff during the phasing maneuvers atapproximately 5000 ft and Mach 0.7. Thephasing maneuvers are performed toensure that the aircraft instrumentationfunctions properly. Because of the natureof the primary experiment, no maneuverswere performed while the aircraft was flyingsupersonically. Therefore, data compari-sons were not made while the aircraft wasexperiencing elevated temperatures. Datawere taken while shock fence temperaturesranged between 100 °F and –40° F.

Results and Discussion

Because data from the forward straingages were similar to that obtained forthe aft strain gages, only data from theforward strain gages are presented in thispaper. Appendix B includes graphs of theaircraft flight conditions, such as Machnumber, dynamic pressure, shock fencetemperature, angle of attack, and angle ofsideslip, which correlate to the straingraphs presented in this paper.

Figure 8(a) shows the strains measuredfrom all of the configurations of the forwardstrain gages while the aircraft completedthree β sweeps during flight 88. All three ofthese constant current loop lead wireconfiguration strains closely follow thestrain measured by the Wheatstone bridge.Figure 8(b) shows the differences betweenthese signals. In this figure, the output fromthe Wheatstone bridge was subtractedfrom the total output of each of the constantcurrent loop lead-wire configurations.Table 1 lists the results.

The five- and three-lead-wire configurationsresponded to the load similarly; meanwhile,the seven-lead-wire configuration gave asmaller response to the load. However

AV PCMmax

E omax---------------------------=

Table 1. Differences between constantcurrent loop and Wheatstone bridge strainmeasurements during phasing maneuvers.

Constant currentloop lead wireconfigurations

Maximum,µin/in

Standarddeviation,

µin/in

7 5.6 1.5

5 7.7 2.1

3 7.3 2.0

10

11

(a) Forward location.

(b) Constant current loop gage responses minus Wheatstone bridge response.

Figure 8. Strain during phasing maneuvers.

Time, sec

Strain,µin/in

20 4030100

150

100

50

0

– 50

970667

5-wire

Point of maximum difference

3-wire

7-wire

Wheatstone bridge

Time, sec

Straindifference,

µin/in

20 4030100

10

5

0

– 5

– 10

970668

5-wire

Point of maximum difference

3-wire

7-wire

because the noise level of the constantcurrent loop is approximately 1.7 µin/in, thisdifference is within the noise level of thedata channel. Before the β sweeps, thedifferences between the constant currentloop measurements and the Wheatstonebridge measurements were smaller. Table 2shows these differences in responses whilethe aircraft is flying straight and level.

Comparing the data from tables 1 and 2reveals that while performing the β sweepsthe differences between the responsesfrom the strain measurement systemsincreased slightly. These changes in thestrain differences and standard deviationsindicate an actual strain differencebecause the constant current loop andWheatstone bridge configured strain gageswere not positioned in exactly the samelocation.

Because the five-lead-wire constantcurrent loop configuration did not requireassumptions about the affects oftemperature on the strain gages and leadwire resistances, it was used as a baselinefor the constant current loop configurations.The differences in the measured strainbetween the constant current loopconfigurations were determined for thephasing maneuvers during flight 88. Agraph of the differences is presented infigure 9. The averages and the standard

Table 2. Differences between constantcurrent loop and Wheatstone bridge strainmeasurements during straight and levelflight.

Constant currentloop lead wireconfigurations

Maximum,µin/in

Standarddeviation,

µin/in

7 3.9 0.93

5 5.0 0.94

3 5.2 0.93

12

Figure 9. Strain differences between the three- and five-wire configurations and the seven-and five-wire configurations during phasing maneuvers.

3-wire

7-wire

Time, sec

Straindifference,

µin/in

20 4030100

2.5

0

– 2.5

– 5.0

– 7.5

970669

deviations of the differences betweenthe five-, three-, and seven-lead-wireconfigurations are given in table 3.

There is a larger strain difference betweenthe seven- and five-lead-wire configurationsthan between the three- and five-lead-wireconfigurations. This difference may occurbecause the seven-lead-wire configurationuses four data amplifiers and four data

channels to record the gage outputs.Conversely, the five- and three-lead-wireconfigurations require two data amplifiersand one data channel each. When theindividual gage responses are combinedinto a total bridge output for the seven-lead-wire configuration, the noise levels of eachof those data channels may be addedtogether. However in this application, thelarger difference between the seven-and five-lead-wire configurations thanbetween the three- and five-lead-wireconfigurations is negligible. Therefore, thethree constant current loop configurationsprovide consistent strain gage outputs.

The total output for the constantcurrent loop seven-lead-wire configurationwas determined mathematically duringpostflight processing of the four individualstrain gage outputs. Figure 10(a) showsthe individual gage responses duringthe flight 88 phasing maneuvers. The

Table 3. Differences between constantcurrent loop strain measurements duringphasing maneuvers.

Average,µin/in

Standarddeviation,

µin/in

3-wire minus5-wire 0.2 0.9

7-wire minus5-wire 4.3 1.0

13

(a) Individual gage response.

Figure 10. Seven-wire configuration during phasing maneuvers.

Time, sec

Gageoutput,counts

20 30 40100

3000

2500

2000

1000

1500

970670

Gage 3

Gage 4

Gage 1

Gage 2

responses from gages 3 and 4, which arelocated next to each other on the outboardsurface of the shock fence, are the exactopposite of the responses from gages 1and 2, which are located in the sameposition on the inboard side.

Figure 10(b) contains the summation ofthese individual gage responses convertedto microstrain. The total bridge response ofthe seven-lead-wire configuration wasdetermined by adding the strain outputs ofgages 1 and 2 and subtracting the strainsof gages 3 and 4. This total was thendivided by 4 because the gains for theindividual gages were set higher thanthose of the five- and three-lead-wireconfigurations and the Wheatstone bridge.

The seven-lead-wire configuration wouldbe used mainly as a rosette to measureindividual gage strain outputs. There is aninherent difficulty with using the seven-lead-wire configuration in this applicationas it was implemented in this experiment.This difficulty was demonstrated duringflight 87. Because of an aircraft emergency

immediately preceding the supersonicportion of the flight, the F-16XL ship 2 wasforced to return to the base at subsonicspeeds. This extended subsonic runcaused the shock fence to experience cold-soaking.

Graphs of the flight conditions includingtemperature are presented in figures B1(a)through B2(e). The strain data from theentire flight is shown in figure 11(a).Differences between the constant currentloop responses and the Wheatstone bridgeare shown in figure 11(b).

The summation of the individual gageresponses does not accurately trackwith the responses from the otherbridges’ configurations. The responses ofgages 3 and 4 from the seven-lead-wireconfiguration went off scale during the flight(fig. 12). A review of the manufacturer’sapparent strain data revealed that thiscondition was caused in part by theapparent strain of the strain gages. Thestrain gages experience a differenttemperature during flight than the strain

14

(b) Combined strain output.

Figure 10. Concluded.

Time, sec

Strain,µin/in

20 30 40100

150

100

50

– 50

0

970671

(a) Forward bridges.

(b) Differences between forward constant current loop and forward Wheatstone bridgeconfigurations.

Figure 11. Strain during flight 87.

160

120

80

40

0

– 40

– 8010000

Strain,µin/in

Time, sec2000 3000 4000

970672

Wheatstone bridge

Lead wires537

20

10

0

– 10

– 20

– 30

– 4010000

Straindifference,

µin/in

Time, sec2000 3000 4000

970673

Lead wires537

gage used as a reference resistance. Thestrain gage which is used as a referenceresistance is located inside the fuselage ofthe aircraft. This reference resistance isused by the signal-conditioning card tosubtract out the unstrained resistance ofthe strain gages in the seven-lead-wireconfiguration (appendix A).

Because it is assumed that the referenceresistance is identical to the resistanceof the strain gages, this temperaturedifference prevents the seven-lead-wireconfiguration from being self-compensatingfor apparent strain when it is used tomeasure individual strain gage responses.When the outputs from the individual strain

15

Figure 12. Individual gage responses from constant current loop seven-wire configurationduring flight 87.

0

400

800

1200

1600

2000

2400

2800

3200

3600

4000

– 40010000

Gageoutput,counts

Time, sec2000 3000 4000

Gage 1

Gage 2

Gage 4

Gage 3

970674

gages are mathematically summed toprovide a total bridge output, this apparentstrain subtracts out. Because neither thefive-lead-wire nor the three-lead-wireconfigurations use the reference resistanceto subtract out the unstrained gageresistances, their outputs are consistent.However, corrections for temperaturewould have to be made to the data from theseven-lead-wire configuration when it isused to measure individual strain gageoutputs.

Constant Current Loop Noise Investigation

Contrary to theoretical expectations, thesignal-to-noise ratio of the constant

current loop signal-conditioning card wasapproximately the same as the Wheatstonebridge during postflight evaluation. Datapresented in figures 13(a) and 13(b) weretaken shortly after the aircraft had landedand while it was at rest with the engines on.These data were used as a baseline for thenoise levels on the airplane. Figure 13(a)shows the noise level of the Wheatstonebridge in volts. The root mean square valueof the Wheatstone bridge noise was0.0137 Vrms. Figure 13(b) shows the noiselevel of gage 1 in the seven-lead-wireconfiguration. Gage 1 had a noise value of0.0267 Vrms. However, the signal outputsof the constant current loop strain gageswere twice as large as those of theWheatstone bridge output.

16

(a) Wheatstone bridge.

(b) Gage 1 of the constant current loop configuration.

Figure 13. Noise level with direct current voltage offset removed.

0 20Samples

Noise,volts

40 60 80 100 120

0

–.1

.1

970675

0 20Samples

Noise,volts

40 60 80 100 120

0

–.1

.1

970676

Figures 14(a) and 14(b) show the signaloutputs of the Wheatstone bridge andgage 1. The Wheatstone bridge signalvalue was 0.5 Vrms, and the constantcurrent loop strain gage had a signal valueof 1.2 Vrms. Therefore, the signal-to-noiseratios for both types of signal-conditioningcards were close to the same value.

Because theoretically, the signal-to-noiseratio of the constant current loop should betwice that of the Wheatstone bridge (ref. 6),an investigation was conducted todetermine the cause of the higher thanexpected noise floor of the constant currentloop signal-conditioning card.

17

(a) Wheatstone bridge.

(b) Gage 1 of the constant current loop seven-wire configuration.

Figure 14. Signal outputs during phasing maneuvers.

Samples

Signaloutput,volts

0 20 40 60 80 100 120

1.0

.8

.6

.4

.2

0

–.2

970677

Samples

Signaloutput,volts

0 20 40 60 80 100 120

2.5

2.0

1.5

1.0

.5

0

–.5

970678

18

Three ground tests were performed duringthe investigation. The output of the dataamplifier was shorted showing that theexcessive noise was being generated bythe constant current loop signal-conditioning card. The resulting noise floorwas acquired on the PCM as counts. Thenoise floors before and after shorting theoutput amplifiers are given in table 4.

Table 4. Constant current loop noisecomparison test 1.

Noisefloor,

Counts

Noise floor with dataamplifier shorted,

Counts

Mean 2146 2062Standard

deviation12 4

The fact that the standard deviationdecreased after the data amplifiers wereshorted indicates that the data amplifierswere generating noise on the signal-conditioning card. However, this findingdid not account for all of the noisebeing generated by the constant currentloop signal-conditioning card. Furtherinvestigation was required to resolve thehigh noise floor problem.

After comparing the Wheatstone bridgesignal-conditioning card to the constantcurrent loop signal-conditioning card, twoconditions were discovered that contributeto the high noise floor of the constantcurrent loop signal-conditioning card. Thefirst difference is that on the constantcurrent loop card all eight amplifier outputreference lines are tied together on theboard. Therefore, only one output pin andwire are available to carry the returncurrent back from the PCM which creates aground loop. To test this theory, sevenoutput amplifiers were removed from theboard, and the output from a single gagemeasurement was taken. This changeallows only the reference line of a singledata channel to be connected to the PCM.Test 2 compared the noise level when alleight output amplifiers are active versuswhen one amplifier is active. The resultsare given in table 5.

The standard deviation for the one amplifierconfiguration shows that when seven of theeight amplifiers were removed, the noisefloor was approximately one-half that of thenormal signal-conditioning card configura-tion. One possible solution to the excessivenoise floor problem would be to reduce thenumber of amplifier output reference linesthat are tied together. In addition, unlikethe Wheatstone bridge signal-conditioningcard, the constant current loop signal-conditioning card does not have a filter afterthe amplifier. It was believed that this lack ofa post-amplification filter was contributingto the high noise floor.

This theory was tested by routing the inputof the data amplifier for a single gagemeasurement from the constant currentloop signal-conditioning card to the input ofthe Wheatstone bridge measurement. Thistechnique caused the constant currentloop single gage output to be processed bythe Wheatstone bridge circuitry whichincludes the data amplifier and an activelow-pass filter. The results from this testare given in table 6.

The noise floor level was reduced toapproximately the same as the Wheatstonebridge when the output from the subtractor

Table 5. Constant current loop noisecomparison test 2.

Eight amplifiersinstalled,Counts

One amplifierinstalled,Counts

Mean 2146 2132

Standarddeviation

12 7

Table 6. Constant current loop noisecomparison test 3.

Single gageoutput,Counts

Single gage outputrouted throughthe Wheatstone

bridge signal path,Counts

Mean 2146 1877

Standarddeviation

12 8

19

was passed through the Wheatstonebridge gain amplifier, active low-pass filter,signal return, and PCM channel. It may beconcluded, therefore, that the noise floorof the constant current loop signal-conditioning card would be significantlyreduced by combining fewer signal returnsand by placing a third-order, active, low-pass filter after the amplifier thusincreasing the signal-to-noise ratio.

Future Testing

Before this constant current loop strainmeasurement system is used on futureflight programs, additional tests should beperformed. Such testing should includeboth laboratory and flight testing. Thelaboratory tests should include placing thetest article under known mechanical andthermal loads and comparing theexperimental data to analytical predictions.Additional flight testing should includeperforming loads maneuvers at supersonicspeeds. This testing would enable thetheoretical benefits of the constant currentloop strain measurement system to be fullyevaluated.

Concluding Remarks

The constant current loop strainmeasurement system provides strainmeasurements without the problem ofreduced sensitivity caused by increasinglead wire resistance which is present withthe Wheatstone bridge method. Additionalbenefits of the constant current loopmethod include the fact that the outputsfrom the constant current loop gages areapproximately twice as large as the outputfrom the Wheatstone bridge. The increasedsensitivity of the constant current loopmethod can mitigate the reduction of gageoutput which results when the excitation

voltage is lowered in the Wheatstonebridge system.

The in-flight evaluation of the constantcurrent loop strain measurement methoddetermined that the strains measured bythe constant current loop compared wellwith the Wheatstone bridge measuredstrains. All three of the constantcurrent loop wire configurations providedconsistent results.

The seven-lead-wire configuration can readindividual gage outputs which is typical ofa rosette. However, corrections must bemade for strain variations experiencedby the individual gages of the seven-wire configuration because of differencesbetween the resistances of the referencestrain gage and the shock fence straingages caused by temperature differencesat the two locations.

The five-lead-wire configuration needsno assumptions made about the leadwire resistances. This configuration isinsensitive to the temperature differencesbetween the reference gage and the straingages.

The benefit of the three-lead-wire configu-ration is that it uses only three lead wires,while the Wheatstone bridge requires four.Like the Wheatstone bridge method, thethree-wire configuration can be susceptibleto measurement errors caused by lead wireresistance changes. However, as long asthe resistances of the two current-carryinglead wires change equally, the output is notaffected. In the case of the Wheatstonebridge system, lead wire resistanceincreases reduce the sensitivity of thesystem to strain. The constant current loopfive- and seven-wire configurations arenot affected by increased lead wireresistances.

20

Two difficulties are currently beingexperienced with the constant current loopflight card system. The first is thatthe signal-to-noise ratio of the constantcurrent loop signal-conditioning card isapproximately the same as the Wheatstonebridge signal-conditioning card becausethe noise floor of the constant current loopsignal-conditioning card is twice as highas the Wheatstone bridge. Constantcurrent loop signal-conditioning cardimprovements, which are needed to reducethe noise level, include minimizing thenumber of reference lines tied together to

the pulse code modulator and placing anactive filter after the data amplifier. Oncethese improvements are made, the signal-to-noise ratio of the constant current loopsystem is expected to be twice that of theWheatstone bridge.

In addition to improving the noise floorof the constant current loop signal-conditioning card, further testing should beperformed before depending on this strainmeasurement system as the primarysource of in-flight strain data.

21

APPENDIX AWHEATSTONE BRIDGE AND

CONSTANT CURRENT LOOP CIRCUITS

Strain gages used in the Wheatstonebridges were configured into two four-active-arm-bending bridges. Figure A1shows a generic diagram of theWheatstone bridge containing four activegages with pairs acting in equal andopposite strain. The output voltage, ,is calculated using the following equation:

(A1)

A detailed derivation of the Wheatstonebridge output is presented in reference 6.

The expected output from the constantcurrent loop is derived using elementarycircuit analysis to determine the voltagedrop across the strain gages. The seven-wire configuration is used to read individualstrain measurements across each gage of

the bending bridge. Figure A2 shows thewiring schematic for the seven-lead-wireconfiguration. For explanation purposes,the output for gage 1 will be derived. Eachgage in the bridge can be represented as afixed resistance and a change in resistancecaused by a strain input. Therefore, thevoltage drop across gage 1 may beexpressed by a constant current, I,multiplied by the sum of the resistance ofgage 1, , and the change in thatresistance, , as shown in thefollowing equation:

(A2)

The and are not included in theequation because, theoretically, theseterms may be neglected because of thehigh input impedance of the amplifier. Inaddition, any influence from the commonmode voltage between the two leads wouldbe minimized by the high common moderejection ratio of the instrumentationamplifiers.

V out

V out E GF( ) ε( )=

Rg1∆Rg1

V g1 I Rg1 ∆Rg1+( )=

Rw 2 Rw 3

22

Figure A1. A Wheatstone bridge configured in a four-active-arm-bending bridge.

Vout

In aircraft instrumentation bay

Signal- conditioning

card

On shock fence

Rg4

Rg3

Rg1

Rg2

E+

–

970679

Figure A2. Constant current loop seven-lead-wire configuration wiring diagram.

Voutg2

Voutg1

Voutg3

Voutg4

Vg1

Vg3

Vg2

Vg4

Signal- conditioning

card

In aircraft instrumentation bayOn shock fence

Rw1

Rw2

Rw3

Rw4

Rw5

Rw6

Rref

Rw7

A

B

C

D

E

F

G

+

–

+

–

+

–

+

– +

–

+

–

+

–

+

–

970680

Rg1 + ∆Rg1

Rg2 + ∆Rg2

Rg3 + ∆Rg3

Rg4 + ∆Rg4

I

To obtain strain measurements, the voltagedrop caused by the unstrained resistance,

must be subtracted out. Thissubtraction is accomplished by placing areference resistor in the current loop.The voltage drop across is subtractedfrom the output of the amplifier at , andthe output is expressed by:

(A3)

The reference resistor is selected so thatits resistance is equal to the gageresistance. The gage and referenceresistance subtract out, and the output issimplified to

(A4)

which is the output voltage due only to thechange in resistance in gage 1.

The expected output from the five-lead-wire configuration is derived in a similarfashion as the seven-wire configuration. Adiagram of the five-wire configuration isshown in figure A3. Lead wire resistancesare denoted by , and the gageresistances are represented by anunstrained resistance plus a change in thatresistance, + , where ∆ resultsfrom a strain input. The and represent the voltage drops across bothhalves of the bridge. The voltage dropsacross and with a constant currentsource I, may be represented as inequations A5 and A6 by

(A5)

(A6)

The , , and may beneglected because of the negligible current

Rg1

RrefRref

V a

V out g1I Rg1 ∆Rg1 Rref–+( )=

V out g1I∆Rg1=

Rw

Rg ∆Rg RgV a V b

V a V b

V a I Rg1 ∆Rg1 Rg2 ∆Rg2+ + +( )=

V b I Rg3 ∆Rg3 Rg4 ∆Rg4+ + +( )=

Rw 2 Rw 3 Rw 4

23

Figure A3. Constant current loop five-lead-wire configuration wiring diagram.

Vout

In aircraft instrumentation bay

Signal- conditioning

card

On shockfence Rw1

Rw2

Rw3

Rw4

Rw5 Rref

B

D

Va

Vb

F

+–

+–

+–

970681

Rg1 + ∆Rg1

Rg2 + ∆Rg2

Rg3 + ∆Rg3

Rg4 + ∆Rg4

I

flow in these leads caused by the high inputimpedance of the instrumentation amplifier.Subtracting from yields thefollowing voltage output:

(A7)

Substituting in equations A5 and A6 for and , respectively, becomes

(A8)

Assuming that all of the gages have thesame initial resistance, simplifies to

(A9)

where the output voltage is the differencebetween the strain gage pairs.

Like the five-lead-wire bridge, the three-wire configuration measures the output offour gages where gage pairs react to equaland opposite strain. The output voltage ofthis card is obtained by calculating and

which are shown in the wiring diagramin figure A4.

Looking at the potential drop across theinput of both amplifiers, the followingequations for and are obtained:

(A10)

(A11)

Unlike the equation for the five-lead-wireconfiguration, and must appearin the equation because no sense leadsare used to read the voltage across both

V b V a

V out V a V b–=

V aV b V out

V out I Rg1 ∆Rg1 Rg2 ∆Rg2+ + +( )[=

– R g 3 ∆ R g 3 R g 4 ∆ R g 4 + + + ( ) ]

V out

V out I ∆Rg1 ∆Rg2+( )[=

– ∆ R g 3 ∆ R g 4 + ( ) ]

V aV b

V a V b

V a I R( w 1 Rg1 ∆Rg1+ +=

+ R g 2 ∆ R g 2 ) +

V b I R( w 3 Rg3 ∆Rg3+ +=

+ R g 4 ∆ R g 4 ) +

Rw 1 Rw 3

24

Figure A4. Constant current loop three-lead-wire configuration wiring diagram.

Vout

In aircraft instrumentation bay

Signal- conditioning

card

On shock fence

Rw1

Rw2

Rw3Rref

A

D

Va

Vb

G

+–

+–

+–

970682

Rg1 + ∆Rg1

Rg2 + ∆Rg2

Rg4 + ∆Rg4

Rg3 + ∆Rg3

I

halves of the bridge. However, maybe neglected because of the high inputimpedance of the amplifier and relativelysmall amount of current drawn into thatlead wire. Subtracting from yields

in the following equation:

(A12)

Substituting in equations A10 and A11, becomes

(A13)

Assuming all of the terms are equal,the voltage drops cancel out, andequation A13 may be simplified to

(A14)

In the special case where all of the leadwires are the same length and the sametemperature, it may be assumed that thelead wire resistances are equal and cancelout, and is expressed by

(A15)

To obtain test data for the three constantcurrent loop configurations, the five- andthree-wire configurations were createdfrom the seven-lead-wire output at theforward bridge location. The five-wireoutput was obtained on the signal-conditioning card by jumpering thevoltages at the B, D, and F locations toother subtractor inputs. These jumperinglocations are shown in figures A2 and A3.The three-wire configuration was createdby jumpering the voltages at the A, D, andG locations to other subtractor inputs.These jumpering locations are shown infigures A2 and A4. The aft constant currentloop gages were configured using the five-lead-wire configuration.

Rw 2

V b V aV out

V out V a V b–=

V out

V out I R(( w 1 Rg1 ∆Rg1+ +=

+ R g 2 ∆ R g 2 ) +

– R ( w 3 R g 3 ∆ R g 3 + +

+ R g 4 ∆ R g 4 ) ) +

RgRg

V out I Rw 1 ∆Rg1 ∆Rg2+ +( )(=

– R w 3 ∆ R g 3 ∆ R g 4 + + ( ) )

V out

V out I ∆Rg1 ∆Rg2+( )(=

– ∆ R g 3 ∆ R g 4 + ( ) )

25

APPENDIX BFLIGHT PROFILES

Figures B1(a) through B1(e) and B2(a)through B2(e) show the Mach number,dynamic pressure, temperature, angle of

attack, and angle of sideslip as a functionof time for the maneuvers described inthis document. These graphs illustratethe flight conditions which F-16XL ship 2was experiencing when the strainmeasurements were taken.

26

(a) Mach number.

(b) Dynamic pressure.

(c) Shock fence temperature.

Figure B1. Flight parameters as a function of time during phasing maneuvers for flight 88.

.78

.76

.74

.72

.70100

Machnumber

Time, sec20 30

970683

570

560

550

540

100

Dynamicpressure,

lb/ft2

Time, sec20 30

970684

78

76

74

72

70

100

Temperature,°F

Time, sec20 30

970685

(d) Angle of attack.

(e) Angle of sideslip.

Figure B1. Concluded.

(a) Mach number.

Figure B2. Flight parameters as a function of time during flight 87.

4.6

4.2

3.8

3.4

3.0

2.6

100

Angle ofattack,

deg

Time, sec20 30

970686

4

2

0

– 2

– 4

– 6

100

Angle ofsideslip,

deg

Time, sec20 30

970687

1.0

.8

.6

.4

.2

10000

Machnumber

Time, sec2000 3000 4000

970688

27

(b) Dynamic pressure.

(c) Shock fence temperature.

(d) Angle of attack.

(e) Angle of sideslip.

Figure B2. Concluded.

600

400

200

0

10000

Dynamicpressure,

lb/ft2

Time, sec2000 3000 4000

970689

80

40

0

– 40

10000

Temperature,°F

Time, sec2000 3000 4000

970690

12

10

8

6

4

2

0

10000

Angle ofattack,

deg

Time, sec2000 3000 4000

970691

56

43210

– 1

– 5– 6

– 3– 2

– 4

10000

Angle ofsideslip,

deg

Time, sec2000 3000 4000

970692

28

References

1Hannah, R. L. and Reed, S. E., ed., StrainGage Users’ Handbook, Elsevier SciencePublishers Ltd., 1992, pp. 67–68.2Anderson, Karl F., “Constant Current LoopImpedance Measuring System That IsImmune to the Effects of ParasiticImpedances,” U.S. patent no. 5,371,469,Dec. 1994.3Anderson, Karl F., A Conversion ofWheatstone Bridge to Current-Loop SignalConditioning for Strain Gages, NASATM-104309,1995.

4Anderson, Bianca T. and Bohn-Meyer,Marta, Overview of Supersonic LaminarFlow Control Research on the F-16XLShips 1 and 2, NASA TM-104257, 1992.

5Dove, Richard C. and Adams, Paul H.,Experimental Stress Analysis and MotionMeasurement, Charles E. Merrill Books,Inc., 1964, pp. 117–124.

6Anderson, Karl F., The Constant CurrentLoop: A New Paradigm for ResistanceSignal Conditioning, NASA TM-104260,1992.

29

REPORT DOCUMENTATION PAGE Form ApprovedOMB No. 0704-0188Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering andmaintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington,VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

4. TITLE AND SUBTITLE 5. FUNDING NUMBERS

6. AUTHOR(S)

8. PERFORMING ORGANIZATION REPORT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING AGENCY REPORT NUMBER

11. SUPPLEMENTARY NOTES

12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

13. ABSTRACT (Maximum 200 words)

14. SUBJECT TERMS 15. NUMBER OF PAGES

16. PRICE CODE

17. SECURITY CLASSIFICATION OF REPORT

18. SECURITY CLASSIFICATION OF THIS PAGE

19. SECURITY CLASSIFICATION OF ABSTRACT

20. LIMITATION OF ABSTRACT

NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)Prescribed by ANSI Std. Z39-18298-102

A Limited In-Flight Evaluation of the Constant Current Loop StrainMeasurement Method

WU 529-31-24

Candida D. Olney and Joseph V. Collura

NASA Dryden Flight Research CenterP.O. Box 273Edwards, California 93523-0273

H-2185

National Aeronautics and Space AdministrationWashington, DC 20546-0001 NASA TM-104331

For many years, the Wheatstone bridge has been used successfully to measure electrical resistance andchanges in that resistance. However, the inherent problem of varying lead wire resistance can cause errors whenthe Wheatstone bridge is used to measure strain in a flight environment. The constant current loop signal-conditioning card was developed to overcome that difficulty. This paper describes a limited evaluation of theconstant current loop strain measurement method as used in the F-16XL ship 2 Supersonic Laminar FlowControl flight project. Several identical strain gages were installed in close proximity on a shock fence whichwas mounted under the left wing of the F-16XL ship 2. Two strain gage bridges were configured using theconstant current loop, and two were configured using the Wheatstone bridge circuitry. Flight data comparing theoutput from the constant current loop configured gages to that of the Wheatstone bridges with respect to signaloutput, error, and noise are given. Results indicate that the constant current loop strain measurement methodenables an increased output, unaffected by lead wire resistance variations, to be obtained from strain gages.

Circuit theory, Constant current loop, F-16XL aircraft, Strain gage, Structuraltestings, Wheatstone bridge

AO3

33

Unclassified Unclassified Unclassified Unlimited

August 1997 Technical Memorandum

Available from the NASA Center for AeroSpace Information, 800 Elkridge Landing Road, Linthicum Heights, MD 21090; (301)621-0390

Presented at the Society of Flight Test Engineers Conference, Orlando, Florida, August 18–22, 1997.

Unclassified—UnlimitedSubject Category 35

Cover pageTitle PageAbstractNomenclatureIntroductionAircraft ConfigurationDescription of Wheatstone Bridge and Constant Curr...Excitation MethodsLead WiresSignal-Conditioning CardsWiring from Signal-Conditioning Card to PCM

Experimental MethodResults and DiscussionConstant Current Loop Noise InvestigationFuture TestingConcluding RemarksAppendix A Wheatstone Bridge and Constant Current Loop Circuits

Appendix B Flight Profiles

ReferencesReport Document Page