Falling Weight Deflectometer Calibration Center and Operational Improvements: Redevelopment of The Calibration Protocol and Equipment

APPENDIX A. FWD CALIBRATION PROTOCOL

INTRODUCTION

This appendix describes the new FWD calibration protocol. It is the basis of the current version of the AASHTO R32-09 procedure, Standard Practice for Calibrating the Load Cell and Deflection Sensors for a Falling Weight Deflectometer.⁽¹⁾ AASHTO R32-09 overrules any discrepancies with appendix A.

In May 2010, additional revisions to AASHTO R32-09 were submitted for review and approved by AASHTO. It is anticipated that the revised procedure will be published as AASHTO R32-11. This appendix is supplemental to the forthcoming AASHTO R32-11, and it provides additional background information about the FWD calibration procedures.

This protocol is written for use with the four types of FWDs currently manufactured or sold in the United States. It is not applicable to the calibration of lightweight deflectometers or cyclic loading pavement testing equipment. Due to differences in design of the four types of FWDs, there are some special considerations for each type, which are described in the annexes.

This protocol does not cover the calibration of other parts of FWDs including temperature probes and distance-measuring instruments. The manufacturers’ recommended procedures should be used to calibrate those devices.

This protocol contains two procedures: annual calibration and monthly calibration. The deflection sensors must be removed from their holders on the FWD and installed in a calibration stand for either procedure.

Annual calibration can either be performed at a FWD calibration center or at a site where the FWD is located. The same equipment and procedures are used at both locations. A certified technician is required to perform the procedure. Annual calibration involves two steps: reference calibration and relative calibration. In reference calibration, the FWD’s deflection and load transducers are calibrated against independently calibrated reference devices. In relative calibration, the deflection sensors are compared to each other.

In monthly calibration, only relative calibration of the deflection sensors is completed at any suitable location using the calibration stand supplied by the FWD manufacturer. A certified technician is not required. Monthly calibration is completed for verification of the accuracy of the deflection sensors and occasionally when a sensor must be replaced. The procedure used for monthly relative calibration is different than for annual relative calibration.

The annual and monthly procedures result in gain factors or dynamic calibration factors which are entered into the FWD software as multipliers. When the FWD raw measurements are multiplied by the gain factors, the result is a value which has been corrected to agree with the calibration instrumentation. It is necessary to have a place to enter the gain factors in the FWD operating system software (also known as the field program) provided by the manufacturer.

Frequency of Calibration

Annual calibration of the FWD load cell and deflection sensors should be performed at least once per year and as soon as possible after a sensor has been replaced.

Monthly calibration should be performed on the deflection sensors at least once per month and immediately after a deflection sensor has been replaced.

WinFWDCal

The calibration protocol as described herein has been automated in a software package named WinFWDCal. It can be used for both annual and monthly calibration. Use of the program is required to carry out the procedure.

ANNUAL CALIBRATION PROCEDURE

For annual calibration, the load cell and deflection sensors are calibrated with the goal of adjusting the accuracy (i.e., systematic error) of the devices to ±0.3 percent or better.

Personnel

Annual calibration requires two people to perform the procedure. One person is the FWD operator, who is responsible for assuring that the FWD is in proper working order for the calibration. During calibration, the operator controls the FWD and removes and replaces the sensors in their holders.

The other person is the calibration operator and is a certified technician. The calibration operator ensures that the calibration equipment is maintained and calibrated as needed. During calibration, this person is responsible for operating the calibration computer and the specialized software used in the calibration of the FWD. In addition, the calibration operator is responsible for providing the documentation of the calibration exercise.

Before beginning a calibration, the FWD operator should present a signed checklist documenting the steps taken in preparation for the calibration, indicating certain preferences concerning the way the calibration should be performed (see annex 1 of appendix A). The FWD operator is responsible for programming the FWD computer to carry out the requested procedure. The FWD operator should provide the history of past calibration results for the FWD.

During calibration, moving the sensors and operating the specialized equipment is a shared responsibility, with the calibration operator having primary control over the calibration equipment. The FWD operator is responsible for transferring the FWD data from the FWD computer to the calibration computer in a format that can be read electronically (including providing a means for the transfers).

After completion of the procedure, the calibration operator should provide the FWD operator with a certificate of calibration that lists the final gain factors for the load cell and each deflection sensor. The FWD operator should enter the final gain factors in the FWD computer and maintain a cumulative history of calibration results in the FWD computer as well as a history of calibration results for the FWD at the calibration center.

Equipment Preparation and Setup Before Calibration

During setup, FWD-specific information will need to be transferred from the FWD operating system (e.g., the FWD field program computer files) to the calibration computer. The annexes in this appendix describe the procedure for obtaining this information for each type of FWD.

FWD

The FWD needs to be in good operating condition prior to performing a calibration. A well-maintained FWD is easier to calibrate and less prone to mechanical and electrical problems during calibration and general use. A checklist to help the FWD operator prepare the FWD for calibration is provided in annex 1. It should be filled out in advance, and a signed copy of the completed checklist needs to be provided to the calibration operator.

Before beginning any calibration work and throughout the entire calibration period, there should not be any data filters turned on in the FWD operating system. The FWD operator is required to verify that all smoothing or filtering has been turned off.

Prior to calibration, the FWD should be warmed up using the standard operating procedure for the particular brand of FWD. After the FWD is moved into position for calibration, it is important that it is level to minimize unwanted vibrations. Trailer-mounted FWDs should also be level to avoid unwanted vibrations.

FWD Drop Sequence

The FWD mass and drop heights/load levels should be set up to produce loads within ±10 percent of the suggested loads shown in table 3. The FWD should be calibrated using three or four load levels. If only three load levels are used, the highest three load levels shown in table 3 should be used.

It is the FWD operator’s prerogative to specify the load levels that will be used for reference calibration. Other load levels may be substituted for those suggested in table 3. The range of loads used should reflect that which FWD normally uses in daily operation. However, in no instance should the combination of static plate load plus maximum dynamic load exceed 24,000 lbf (106 kN). This limitation is required to protect the reference load cell and the concrete pavement used in the calibration procedure.

During setup, the minimum number of drops at each load level will be determined by WinFWDCal based on the deflection response of the concrete pavement or test pad. More than the minimum number of drops may be used without exceeding 10 drops per load level. The FWD operator should program the drop sequence in the FWD computer, progressing from the lowest to the highest load level. The same number of drops should be used at each load level, and the same drop sequence should be used for both load and deflection sensor calibration.

The software may inform users that the deflections are either too large or too small to satisfy the precision requirements for reference calibration. The vertical accelerations are also checked to ensure that they do not exceed ±5 g. If the deflection or the acceleration is too large, the FWD should be moved further away from the sensor stand. If the deflection is too small and if three load levels are used, users should try four load levels. If the deflection is still too small and the FWD cannot be moved closer to the sensor stand, users should try a sequence of higher load levels, if possible. If this does not solve the problem, users should find an alternative location to calibrate the deflection sensors.

FWD Calibration Equipment

Table 4 provides information on the equipment needed to perform the load cell and deflection sensor calibration. Detailed information for all components is found in appendix E of this report.

Both reference and relative calibration of all sensors should be performed on concrete pavement. The concrete floor area, or optional concrete test pad, should be in good condition with little or no cracking. The ball-joint base should be attached firmly to the concrete with two anchor bolts to hold the sensor stand in direct contact with the concrete. Additionally, the sensor stand should be clamped tightly in the base. Slippage, rocking, or vibration between the stand and the base or between the base and the concrete is not allowed. The ball-joint should rotate with slight friction.

Reference Load Cell Calibration

The reference load cell should be calibrated at least once per year in accordance with the AASHTO R33-03 procedure (see annex 7 of appendix A).⁽¹⁰⁾

Accelerometer Calibration

The accelerometer is calibrated in Earth’s gravity prior to use for relative calibration to determine the daily response of the accelerometer. This calibration should be repeated after 8 h have elapsed.

The accelerometer, mounted in an aluminum box, should be calibrated using the calibration platform. The platform should be carefully adjusted using the bubble level to assure that the accelerometer is aligned with Earth’s gravity. The accelerometer needs to be calibrated in both ±1 g fields by inverting the box briefly. Care must be taken to avoid dropping the accelerometer during the calibration process because the shock may cause damage.

WinFWDCal will guide the calibration operator through the accelerometer calibration procedure to calculate the calibration coefficients. The accelerometer box should not be in the -1 g orientation for more than 30 s during the calibration process to minimize the effect of hysteresis on the readings. If it is inverted for a longer period of time, the accelerometer calibration process needs to be stopped. Next, the box should be placed upright in a +1 g field for at least twice as long as it was inverted (up to a maximum of 24 h) to return it to equilibrium. For example, if the box is upside down for 1 min, it should be allowed to equilibrate for at least 2 min before repeating the calibration of the accelerometer.

The accelerometer calibration is slightly temperature sensitive. As a result, it is important that the accelerometer is calibrated shortly before its use. Temperature is monitored continuously by WinFWDCal. The program will alert the calibration operator if the temperature changes by more than ±18 °F (±10 °C). If this occurs, the accelerometer needs to be recalibrated.

FWD Calibration Procedure

It is recommended to complete that deflection sensor calibration first, followed by the load cell calibration. WinFWDCal should be used to collect data and make the associated QA checks.

Table 5 shows the calibration data reporting requirements and the sources of the data. Most of the data are read electronically by the calibration computer from files copied from the FWD operating system. All of the data can be entered or updated manually using WinFWDCal.

Deflection Sensor Calibration

Deflection sensor reference calibration consists of two trials where all of the sensors are calibrated simultaneously in a special stand (see appendix F). The position of the sensors in the stand is inverted between the trials. This is followed by two relative calibration trials using the same stand, where the sensors are inverted once more. Spare deflection sensors should not be calibrated unless they have separate dedicated signal-conditioning channels in the FWD microprocessor.

For reference and relative calibration, the sensor stand should be manually held and kept vertical (as indicated by the bubble level) while data are being collected so the accelerometer box will be correctly aligned with Earth’s gravity. The sensor stand should be supported by the rest stop between trials.

Reference Calibration

Deflection sensor reference calibration consists of performing at least two trials in which all of the sensors are calibrated simultaneously in the sensor stand. The position of the sensors in the stand is inverted between each trial. Spare deflection sensors should not be calibrated unless the FWD has separate dedicated signal-conditioning channels.

The deflection sensors should be placed in the sensor stand and centered on the reference accelerometer. WinFWDCal displays a diagram showing how to arrange the sensors in the stand (see table 8 and table 9 in appendix B).

The drop sequence for the first reference calibration trial should be performed. The operators should review and accept or reject the data for each drop. WinFWDCal will graphically display the deflection time history data after each drop.

Excessive drift from the baseline shown on the time history graph should be cause for rejection of a drop. The software will alert the user when drift is excessive.

At the conclusion of each trial, the user should transfer the data electronically from the FWD to the calibration computer and review the results. For each sensor, WinFWDCal will regress the FWD output (independent variable) versus the reference deflection sensor (dependent variable) forced through zero. The slope of the regression line for each sensor, when multiplied times the initial gain factor, gives the reference gain factor. The slope for an individual sensor is acceptable if its standard error does not exceed 0.0020. The trial is acceptable if the standard errors for all sensors do not exceed 0.0020.

If the first trial is acceptable, the user should continue with the second trial. The sensors in the stand should be inverted before the second trial according to the diagram displayed by WinFWDCal. If any trial is not acceptable, the data should be rejected and the trial is repeated. The user should investigate the reason why the standard error exceeded 0.0020 and correct it, if possible, before repeating the procedure. The user must verify that all sensors are held firmly in the stand.

Interim Gain Factor Acceptance Criteria

After two reference calibration trials have been accepted, WinFWDCal will calculate the average reference gain factor for each sensor and display the results as the interim gain factors (one gain factor for each deflection sensor).

WinFWDCal will calculate the difference in the reference gain factors for each sensor between the two trials. If the difference for each sensor is no more than 0.005, then the reference calibration test is complete. If any of the differences is greater than 0.005, two additional trials should be performed. The user should accept the results and continue but note if any of the differences is more than 0.005. The average of the reference gain factors for all trials for each sensor should be reported as the interim gain factors.

Relative Calibration

Relative calibration is followed by reference calibration and uses the same sensor stand. Two trials are performed. For each trial, 40 drops are applied from the highest drop height used in reference calibration. The sensors should not be repositioned in the sensor stand before the first trial.

WinFWDCal will adjust the FWD data collected in the relative calibration using the interim calibration factors internally. The user should not enter the interim factors in the FWD operating program before performing relative calibration.

To begin, the user should perform the drop sequence for the first relative calibration trial. At the conclusion of each trial, the data should be transferred electronically from the FWD to the calibration computer, and the results should be reviewed. For each sensor, WinFWDCal will calculate the means ratio. The means ratio multiplied times the interim gain factor gives the relative gain factor.

WinFWDCal performs an ANOVA test for the data and reports the standard error. The trial is acceptable if the standard error is not more than 0.12 mil (3 μm) and there are no extreme outliers in the data.

WinFWDCal will display a plot of the data for the 40 drops. The graph should be scanned to detect extreme outliers (i.e., due to a loose sensor in the stand). An extreme outlier would appear substantially outside the normal range of the deflection data.

If the standard error is greater than 0.12 mil (3 μm) or if there are extreme outliers in the data, the first trial is not acceptable, the data should be rejected, and the trial should be repeated. Note that the sensors in the stand should not be repositioned. The reason why the data are unacceptable should be investigated and corrected, if possible, before the procedure is repeated. However, if the first trial is acceptable, the user should continue with the second trial. The sensors in the stand should be inverted before the second trial according to the diagram displayed by WinFWDCal. After two trials have been accepted, WinFWDCal will calculate the average relative gain factor for each sensor and report the results as the final gain factors, completing the deflection sensor calibration procedure.

Note that if two acceptable relative calibration trials cannot be obtained after performing four trials, no further effort should be made to calibrate the deflection sensors.

Load Cell Reference Calibration

If the reference load cell has not been calibrated within the past 12 months, it should be recalibrated in accordance with AASHTO R33-03.⁽¹⁰⁾

Reference load cell calibration consists of at least two trials. The FWD load plate should not be raised at any time during the procedure. To begin, users should perform the drop sequence for the first reference calibration trial. After, they should review and accept or reject the data for each drop. WinFWDCal will graphically display the deflection time data after each drop.

At the conclusion of each trial, the data should be transferred electronically from the FWD to the calibration computer, and the results should be reviewed. WinFWDCal will regress the FWD output (independent variable) versus the reference load cell (dependent variable) forced through zero. The slope of the regression line for each sensor, when multiplied times the initial gain factor, gives the reference gain factor. The slope for the trial is acceptable if its standard error is not more than 0.0020.

If the first trial is acceptable, the user should continue with the second trial. However, if the first trial is not acceptable, the data should be rejected, and the trial should be repeated. The user should investigate the reason why the standard error exceeded 0.0020 and should correct it, if possible, before repeating the procedure. The load cell should sit squarely under the FWD load plate and squarely on the concrete.

Gain Factor Acceptance Criteria

After two trials have been accepted, WinFWDCal will calculate the average reference gain factor and report the results as the final gain factor. If the range of the two reference gain factors does not exceed 0.003, then the final gain factor should be accepted, completing the load cell calibration procedure.

However, if the results of the first two trials are outside the acceptable range, a third reference calibration trial should be performed. If the standard deviation of the gain factors for three acceptable trials does not exceed 0.003, then the results of the three trials should be averaged and reported as the final gain factor for the load cell, thus completing the load cell calibration procedure.

If the standard deviation of the three trials exceeds 0.003, the reference load cell calibration procedure should be repeated, yielding a fourth reference gain factor. If the standard deviation of all calibrations (four acceptable trials) does not exceed 0.003, the average of all four results should be reported as the final gain factor for the load cell, and the load cell calibration procedure is complete.

If acceptable results cannot be obtained after performing four trials, no further effort should be made to calibrate the load cell.

Evaluation and Acceptance of Final Results

WinFWDCal will perform the needed calculations. The data should be evaluated as follows.

1. The final gain factors from this calibration should be compared to the corresponding gain factors from the previous calibration (i.e., the initial gain factors). There should not be more than a 1 percent difference for each individual deflection sensor and for the load cell. If this criterion is satisfied, the final gain factor for the sensor is accepted. However, if this criterion is not satisfied for a sensor, it should be evaluated according to the next criterion.
   
   
2. The final gain factor for the sensor should fall within the range of 0.980 to 1.020. If this criterion is satisfied, the final gain factor for the sensor should be accepted. However, if this criterion is not satisfied for the sensor, it should be evaluated according to the next criterion.
   
   
3. If a historical record of previous calibrations according to this procedure is available for the sensor for 4 years or more and there are at least three previous calibration results over this time period, then the best fit time rate of change of the final gain factor for the sensor should not exceed 0.3 percent per year. If this criterion is satisfied, the final gain factor for the sensor is accepted.

Certificate of Calibration

If the final calibration results meet the acceptance criteria for all sensors according to one or more of the three criteria listed above, the calibration operator should provide the FWD operator with a certificate of calibration listing the final gain factors for the load cell and each deflection sensor. These factors should be entered into the FWD computer. An output file in PDDX format is provided by WinFWDCal to facilitate electronic data transfer.

Report and Retention of Data

The FWD operator should be provided with an electronic copy and a hardcopy of the calibration results. The FWD calibration report should consist of the following:

• The certificate of calibration or a summary page of the results if the calibration was not successful.
  
  
• An electronic copy of the complete record of the annual calibration including the data transferred from the FWD.
  
  
• A pre-calibration checklist.
  
  
• Any other records should be annotated with the FWD unit identification (e.g., manufacturer's serial number or agency ID) and the calibration date.

Calibration records should be retained by the calibration center for at least 5 years.

MONTHLY CALIBRATION PROCEDURE

Monthly calibration serves two purposes. First, it is a means to verify that the deflection sensors are functioning properly and consistently. Second, it can also be used to replace a damaged sensor, providing a short-term gain factor for the replacement sensor until an annual calibration can be performed.

Monthly calibration uses a calibration stand provided by the FWD manufacturer. The deflection sensors are stacked vertically in the stand so that all of the sensors are subjected to the same pavement deflection. Position in the stand may have an effect on the deflection readings. To compensate for this, the sensors are rotated through all positions in the stand. This rotation procedure is different from the relative calibration procedure performed for annual calibration.

Relative calibration relies on collecting a large amount of data that can be averaged to reduce the significance of random measurement errors. Deflections in excess of 20 mil (500 μm) are needed for the results to be accurate. WinFWDCal does the statistical data analysis to compute adjustment ratios and final gain factors from the data.

Since a large number of drops are involved, the properties of the pavement materials may change due to compaction or liquefaction during the procedure. However, all of the sensors are equally affected, and as long as the effect is not too extreme, the adjustment ratios are still accurate.

Some FWDs may have less than seven or more than nine active deflection sensors. If they do, these procedures should be modified to simultaneously calibrate the actual number of active sensors.

Equipment Preparation

FWD

The FWD should be in good operating condition prior to calibration. A well-maintained FWD is easier to calibrate and less prone to mechanical and electrical problems during calibration and during general use. A checklist to help the FWD operator prepare the FWD for calibration is provided in annex 1.

Before beginning any calibration work and throughout the entire calibration period, there should not be any data filters in operation in the FWD operating system. The FWD operator must verify that all smoothing or filtering has been turned off. Additionally, prior to calibration, the FWD should be warmed up using the standard operating procedure for the particular FWD brand.

Other Equipment

Additional calibration equipment includes the FWD calibration stand provided by the manufacturer.

Procedure Overview

Replicate deflection readings should be taken with the sensors assembled in the calibration stand. With the sensors in a particular position in the stand, two unrecorded seating drops followed by five recorded drops constitute a set. The deflection sensors should be rotated through the various positions in the calibration stand in a prescribed way. The rotation procedure is shown in table 6 for a nine-sensor system and table 7 for a seven-sensor system. The total number of sets of data is equal to the number of sensors on the FWD.

The test point (the location where the FWD load plate is positioned) should be “conditioned” before beginning the calibration procedure to reduce the significance of the set in the data analysis. The warm-up drops should be used for this purpose.

Monthly Calibration of the Deflection Sensors

Follow the WinFWDCal on-screen instructions for setup of the procedure as follows:

1. Remove all the deflection sensors from their holders on the FWD. For an FWD with n sensors, number the sensors from D1 to Dn with respect to their normal position on the FWD. The center position is sensor number "1" on the Dynatest^®, Carl Bro, and JILS FWDs and sensor number "0" on the KUAB FWD. In either case, the center sensor should be defined as D1 for this procedure.
   
   
2. Label the levels on the sensor stand from "A" to "G" or "I" as appropriate. The top level is labeled "A."
   
   
3. Position the deflection sensors in the stand for the first set as shown in table 6 or table 7.
   
   
4. Position the sensor stand vertically. Mark the location where it rests so that it can be relocated precisely on the same spot. This should be performed by gluing a washer to the pavement or by chipping a small divot in the pavement with a chisel or a screwdriver.
   
   
5. Select the FWD drop height and the distance from the loading plate to the sensor stand to yield deflections near 20 ±4 mil (500 ±100 μm). If deflections in this range cannot be achieved, choose another location. A concrete pavement on a relatively weak subgrade will usually yield the required deflection.
   
   
6. Lower the FWD loading plate. If the FWD operating system allows, do not raise the loading plate or move the FWD during relative calibration testing.
   
   
7. Make two seating drops (no data recorded) for each set followed by five replicate drops (for which data should be recorded).
   
   
8. Rotate the sensors in the stand at the end of each set as shown in table 6 and table 7. The general progression is for the sensors to move from the bottom toward the top of the stand.

   Note: The rotation must be performed as prescribed for the data analysis in WinFWDCal to work properly. If the direction of rotation is reversed, the calculations will be incorrect.

9. Record 45 drops for a 9-sensor FWD and 35 drops for a 7-sensor FWD.

Monthly Calibration Data Analysis

The data file should be transferred electronically from the FWD to WinFWDCal for analysis. Options provided in the software indicate whether a normal data analysis or a special analysis is required for sensor replacement.

Adjustment of Gain Factors—Normal Analysis

WinFWDCal will report the adjustment ratios and the gain factors for each deflection sensor. Since sensor replacement is not involved, the adjustment of the gain factors in the FWD operating system should be made only when those changes are both significant and verified to be necessary. The following guidelines should be used to evaluate the need to adjust the gain factors:

• If all of the adjustment ratios fall between 0.997 and 1.003, then they are not statistically significantly different from a ratio of 1.000. The calculated adjustments are trivial, the calibration test is complete, and no change of the gain factors should be made.
  
  
• When the adjustment ratios for one or more sensors fall outside of the range 0.997 to 1.003, a second relative calibration trial should be performed. If the gain factors for each sensor in both trials agree within 0.003, the need for the adjustment has been verified. The gains should be adjusted for all sensors. The gain factors for each sensor for the two trials should be averaged and entered in the FWD operating system, and the calibration test is complete.
  
  
• If the difference in the gains for the two trials for one or more sensors is greater than 0.003, two additional trials should be performed. If the second attempt also exceeds the allowable difference, the annual calibration procedure should be performed as soon as possible.

Adjustment of Gain Factors—Sensor Replacement

When replacing a damaged deflection sensor, the monthly calibration procedure should be used to determine a gain factor for the replacement sensor. The calculations are performed by WinFWDCal, and a gain factor is reported only for the replacement sensor.

Two relative calibration trials should be performed. If the two gain factors agree within 0.003, then the gain factors for the two trials should be averaged and entered in the FWD operating system, and the calibration test is complete. However, if the difference is greater than 0.003, two more relative calibration trials should be performed. WinFWDCal will report the average gain factor from the four trials and the standard deviation. If the standard deviation is not more than 0.003, then the average gain factor should be entered in the FWD operating system, and the calibration test is complete. If the standard deviation is more than 0.003, no further effort should be made to calibrate the replacement sensor, and the annual calibration procedure should be performed as soon as possible.

Report

The relative calibration report consists of all printouts from WinFWDCal annotated as necessary to explain any problems which may have occurred.

ANNEX 1. PRE-CALIBRATION CHECKLIST

Fill out and bring this checklist with you to the calibration center. Your signature below indicates that you have met all of the pre-calibration requirements.

ANNEX 2. SPECIAL PROCEDURES FOR CALIBRATING A DYNATEST^® FWD

1. The FWD operating system (i.e., the field program) should be programmed to provide the output from the FWD in the F10, F20, or F25 format. All file extensions should be Fnn, where nn is the version number of the FWD field program. This format is needed so that PDDXconvert can read the FWD file for WinFWDCal.
   
   
2. Save an F25 file in addition to the Microsoft Access^® database file for use with PDDXconvert for FWDs using Dynatest^®’s FWDWin as the operating system.
   
   
3. Warm up the FWD off the calibration test pad. Save the data in a file named “Warmup.Fnn,” where nn is the version number of the FWD field program.
   
   Note: During the setup procedure in WinFWDCal, the Warmup.Fnn file will be read electronically to get information on sensor serial numbers, initial gains, etc.
   
   
4. Power the FWD with a battery charger. Without a charger, the FWD electrical system may not have enough power for a full calibration. Charging the FWD should be done in accordance with the manufacturer’s recommendations. Some chargers may cause electrical noise on the time history signal during reference calibration.
   
   
5. Attach the trailer to the tow vehicle during calibration for trailer-mounted FWDs.
   
   
6. For the Dynatest^® FWD, it is possible to be within the tolerance for the highest load and have the drop height set too high. Before placing the reference load cell under the load plate and with the mass positioned at drop height four (the highest position), users must verify that there will be no interference between the catch mechanism and the brace between the two columns that surround the cylinders that raise and lower the load plate. The reference load cell is nearly 3.875 inches (99 mm) high, so the catch mechanism will be that much higher during load calibration. If the clearance is too small, the target for the fourth drop height should be repositioned to provide the needed clearance.
   
   
7. Remove the removable masses to get the peak load from the highest drop height into the range of 16,000 lbf (71kN) ±10 percent for the Dynatest^® HWD. This will avoid exceeding the capacity of the reference load cell and avoid damaging the test pad due to excessive deflections.
   
   
8. Clean out all stones that are embedded in the rubber pad under the load plate. The FWD operator should complete this cleaning before arrival at the calibration center.
   
   
9. Assemble the calibration stand hardware before FWD arrives. The knurled knob should be placed above the shelf in the calibration stand (see figure 33). The knurled knob should only be tightened by hand. Do not use a wrench to attach the hardware.
   
   
10. Clean the bottom of the geophones and remove any small particles that will interfere with proper contact of the magnetic mounts. Emery paper does a good job of this. The FWD operator should complete this cleaning before arrival at the calibration center.
    
    
11. Place a geophone on the knurled knob in the stand and try to rock it lightly back and forth. If there is any motion, reclean the bottom of the sensor. Each geophone must be securely fastened in the stand.
    
    
12. Place a wooden block under the FWD’s raise-lower bar during load cell calibration to reduce the noise and vibration.
    
    
13. Ensure that the air in the large hydraulic cylinder has been bled out. Failure to bleed the air from the hydraulic system will result in vibration when the catch is released. This will delay the calibration due to excess noise messages. The FWD operator should perform the bleeding before arrival at the calibration center.
    
    
14. Verify that the new calibration gains have been entered correctly into the FWD field program by printing a summary of the relative gains. The relative gain is the value that is determined from the calibration. Do not change the reference gains because those are determined by the manufacturer.

Figure 33. Photo. Attachment of a Dynatest^® geophone in the stand.

ANNEX 3. SPECIAL PROCEDURES FOR CALIBRATING A JILS FWD

1. The FWD operating system (i.e., the field program) should be programmed to provide output with DAT file extensions. This format is needed so that PDDXconvert can read the FWD file for the WinFWDCal software.
   
   
2. The JILS calibration operating system software will require the FWD operator to release each drop manually. It is used during reference calibration. JTEST is used during relative calibration.
   
   
3. Warm up the FWD off the calibration test pad. Save the data in a file named “WARMUP.DAT.”
   
   Note: During the setup procedure in WinFWDCal, there are three files from the FWD computer that need to be read electronically in order to obtain information about the FWD sensor serial numbers and gains. The WARMUP.DAT file contains the deflection sensor serial numbers, the SETUP.PAR file contains the deflection sensor current gains, and the JILS.CFG file contains the load cell current gain.
   
   
4. Type the sensor gains or serial numbers manually into the WinFWDCal setup screen if the sensor gains or serial numbers cannot be read electronically.
   
   
5. Read the three configuration files and verify that each serial number is different. If they are not, rename the sensor numbers on the WinFWDCal setup screen with a number matching the sensor position on the FWD. For example, if all of the sensor serial numbers are “CHOP,” rename them “CHOPn” where n is the number of the sensor starting at 1 for the center sensor at the load plate.
   
   
6. Attach the trailer to the tow vehicle during calibration for trailer-mounted FWDs.
   
   
7. Shim under the wheels as needed to assure both vehicles are level. The tongue wheel can also be used to help level the trailer. Both the tow vehicle and the trailer must be level when the FWD load plate is on the floor.
   
   
8. Check that the truck-mounted FWDs are level when the load plate is down. Check this on the chassis with a carpenter's level. Shim under the wheels if necessary to get the vehicle level.
   
   
9. Check that the load plate is squarely seated on the reference load cell and on the concrete test pad. The JILS load plate does not have a swivel, so the load plate may only be making partial contact. When good contact is achieved, it will not be possible to slide a piece of copy paper under the load plate more than a fraction of an inch (a few millimeters) at any point around the perimeter of the plate. Apply several seating drops (no data recorded) before making this check.
   
   
10. Remove the geophone sensor bar attached to the back of the FWD to allow the FWD to get close to the ball-joint anchor and the geophone stand.
    
    
11. Attach each geophone in the calibration stand with the knurled knob below the shelf as shown in figure 34. The knurled knob should only be tightened by hand. Do not use a wrench to attach the hardware.
    
    
12. Check that the quick connectors for the geophones are working properly. If they are not making good electrical contact, the FWD may provide intermittent noise or a sensor may fail to respond at all.
    
    
13. Power the FWD with a battery charger. Without a charger, the FWD electrical system may not have enough power for a full calibration. Charging the FWD should be performed in accordance with the manufacturer’s recommendations. Some chargers may cause electrical noise on the time history signal during reference calibration.
    
    
14. Verify that the new calibration gains have been entered correctly into the FWD field program by printing a copy of the WARMUP.DAT, SETUP.PAR, and JILS.CFG files. The deflection sensor serial numbers are in the line that begins “Sensors:” near the top of the file. The geophone gains are in the SETUP.PAR file in the line below the text “# Sensor channels EU conversion factors:” The load cell gain is in the JILS.CFG file.

Figure 34. Photo. Attachment of a JILS geophone in the stand.

ANNEX 4. SPECIAL PROCEDURES FOR CALIBRATING A KUAB FWD

1. The FWD operating system (i.e., the field program) should be programmed to provide output of with FWD file extensions. This format is needed so that PDDXconvert can read the FWD file for the WinFWDCal software.
   
   
2. The FWD operating system software may require the FWD operator to release each drop manually. It may not be possible to program the desired drop sequence. In order to provide the pause needed after each drop during reference calibrations, the KUAB operating system software should be set up to perform manual drops.
   
   
3. During the setup procedure in WinFWDCal, the FVO.INI file from the FWD computer will be read electronically to get information on sensor serial numbers, initial gains, etc.
   
   Note: If the sensor initial gains or serial numbers cannot be read electronically, type the values into the WinFWDCal setup screen.
   
   
4. Power the FWD with a battery charger. Without a charger, the FWD electrical system may not have enough power for a full calibration. Charging the FWD should be done in accordance with the manufacturer’s recommendations. Some chargers may cause electrical noise on the time history signal during reference calibration.
   
   
5. Attach the FWD trailer to the tow vehicle during calibration.
   
   
6. Conduct a static calibration of the seismometers before the reference calibration procedure is performed. The KUAB software will automatically file the static calibration factors.
   
   
7. It will only be possible to calibrate the 11.7-inch (300-mm) load cell on the KUAB. If the KUAB is outfitted with load cells with a diameter larger or smaller than 11.7 inches (300 mm), they should not be calibrated using the 11.7-inch (300-mm)-diameter reference load cell because the results would not be accurate.
   
   
8. Locate the FWD load plate as close as possible to the ball-joint anchor. This may require aligning the FWD beside the seismometer stand. Move the FWD into position before putting the seismometer stand in the ball-joint to avoid possible damage to the stand. The FWD should not touch the calibration stand during calibration.
   
   
9. Use the special two-column seismometer stand for deflection sensor calibration for KUAB FWDs with seismometers. The sensors are mounted on standoffs that are the same size as the ones on the FWD. After the set screw has been tightened, the seismometer should be firmly attached, and it should not wobble or rock on the standoff. Do not over tighten the set screw because the hole may be stripped. Place the geophones in the calibration stand as diagrammed by WinFWDCal.
   
   Note: The deflection sensor that is mounted through the load plate (i.e., the center sensor) is called sensor number zero on KUAB, but it is sensor number 1 as far as WinFWDCal is concerned.
   
   
10. Keep the stand aligned perpendicular to the load pulse from FWD during calibration of the seismometers. The easiest way to do this is to stand facing the load plate of FWD.
    
    
11. Attach each geophone in the calibration stand with the knurled knob below the shelf after removing the deflection sensors from the FWD as shown in figure 34. KUABs outfitted with geophones use the same one-column calibration stand used for Dynatest^® geophones. The knurled knob should only be tightened by hand. Do not use a wrench to attach the hardware.
    
    
12. Place the geophones in the calibration stand as diagrammed by WinFWDCal. A magnetic collar holds the KUAB geophone to the calibration stand as shown in figure 35 and figure 36. A special magnetic cup holds the collar to the stand.
    
    
13. Enter the final gain factors for the deflection sensors in the sharpcalfa variable within the applicable [sensor#] section of the FVO.INI file. “#” refers to the sensor number with the numbering starting at zero.
    
    
14. Enter the final gain factor for the load cell in the sharpcalfa variable within the [loadplate] section of the FVO.INI file, if available. If this section is not in the file, the load cell testing should be considered to be calibration verification only.
    
    
15. Verify that the new calibration gains have been entered correctly into the FWD field program by printing a copy of the FVO.INI file.

Figure 35. Photo. KUAB geophone parts.

Figure 36. Photo. Attachment of a KUAB geophone in the stand.

ANNEX 5. SPECIAL PROCEDURES FOR CALIBRATING A CARL BRO FWD

1. The FWD operating system (i.e., the field program) should be programmed to provide output with FWD file extensions. This format is needed so that PDDXconvert can read the FWD file for WinFWDCal.
   
   
2. The FWD operating system software may require the FWD operator to release each drop manually. It may not be possible to program the desired drop sequence.
   
   In order to provide the pause needed after each drop during reference calibration, the operating system software can be set up to perform manual drops without raising the plate only after performing the seating drops (no data recorded) and at least one recorded drop. The FWD operator should prepare a list of the drops needed ahead of time and be prepared to reject the first drop after the seating drops.
   
   
3. In order to obtain the sensor serial numbers and the initial gains, use the RoSy^® software from Carl Bro to create a text file containing these values. The file should be named “HW_INI.TXT.”
   
   Note: During the setup procedure in WinFWDCal, the HW_INI.TXT file will be read electronically to get information on sensor serial numbers, initial gains, etc.
   
   
4. Type the values of the sensor initial gains or serial numbers into the WinFWDCal setup screen if they cannot be read electronically.
   
   
5. Attach the trailer to the tow vehicle during calibration for trailer-mounted FWDs.
   
   
6. The Carl Bro FWD uses a generator on the trailer to provide the electrical power for operations. If possible, warm up the FWD outside and fully charge the batteries prior to moving the FWD indoors. This will allow the calibration to proceed without turning on the generator any more than necessary.
   
   
7. Locate the FWD load plate as close as possible to the ball-joint anchor. This may require placing the geophone stand inside the rear of the FWD frame. Move the FWD into position before putting the geophone stand in the ball-joint to avoid possible damage to the stand. The FWD should not touch the calibration stand during calibration.
   
   
8. Remove the deflection sensors from the FWD and attach each geophone in the calibration stand with the metric knurled knob below the shelf as shown in figure 37. The knurled knob should only be tightened by hand. Do not use a wrench to attach the hardware.
   
   
9. The Carl Bro operating system software does not have a specific place for the new gain factors, so the final gain factors must be multiplied by hand and input into the FWD operational software in the correct place. Follow the manufacturer’s recommendations for this step.
   
   
10. Verify that the new calibration gains have been entered correctly into the FWD field program by printing a copy of the updated HW_INI.TXT file. Use the RoSy^® software from Carl Bro to create this text file.

Figure 37. Photo. Attachment of a Carl Bro geophone in the stand.

ANNEX 6. SPECIAL PROCEDURES FOR ON-SITE CALIBRATIONS

The calibration protocol allows users to perform annual calibrations at locations other than a calibration center. The equipment is highly portable, and it may be advantageous to perform the calibration where the FWD is located rather than shipping the FWD to a center. This procedure is called on-site calibration because it is performed at a site preferred by the FWD owner.

The calibration procedure outlined in appendix A should be used without exception. It must be carried out by a certified calibration center operator as required in AASHTO R 32-09.⁽¹⁾

A special test pad is not required to be constructed for on-site calibrations. In this regard, AASHTO R 32-09 indicates the following:⁽¹⁾

While an isolated test pad is recommended, it is not required, provided that all other facilities requirements, especially the minimum slab deflection requirement and sufficient slab damping, are achieved. The slab dimensions (4 by 5 m) are suggested, and other dimensions may be satisfactory … In general, a concrete pavement on a relatively weak subgrade will yield the required deflection amplitude.

As previously noted in this report, the recommended peak deflection is 20 mil (500 µm), and the minimum recommended deflection is 12 mil (300µm). Peak deflection is a useful way to identify suitable locations for performing calibrations. However, deflection is not the limiting criterion. WinFWDCal checks for the following two conditions:

• The maximum acceleration must not exceed 5 g.
  
  
• The maximum number of replicate drops must not exceed 10 drops at each load level.

The difficult part of on-site calibration is finding a suitable place to mount the ball-joint anchor and the calibration stand. Figure 38 shows a calibration being performed at the Hawaii Department of Transportation in a breezeway between two buildings. The concrete pavement at this location was relatively thin, and when the FWD was close to the calibration stand, the maximum acceleration exceeded 5 g. To overcome this problem, the FWD trailer was moved about 24 inches (600 mm) in front of the stand, thereby reducing both the peak deflections and maximum accelerations.

Figure 38. Photo. On-site FWD calibration at Hawaii Department of Transportation Materials Lab.

Note that the FWD must be level during the calibration. It is also important to have the calibration equipment protected from the weather at all times. To assess the suitability of a location for FWD calibration, it may be helpful to use the Facilities Review Form found in figure 45 in appendix D.

A procedure involving an FWD and one deflection sensor is used to find a place to install the ball-joint anchor. The sensor is removed from its holder and held in place by hand on the pavement. A firm downward pressure is required to keep the sensor in contact with the pavement. A concrete pavement works better than an asphalt pavement because it spreads the deflection basin over a wider area. The procedure is as follows:

1. Position the deflection sensor 3.9 inches (100 mm) from a joint in the concrete. Position the FWD so the rear of the load plate is 30 inches (750 mm) ahead of the deflection sensor (34 inches (850 mm) from the slab joint). If there is no pavement joint, draw a line across the pavement that can be used to measure from.
   
   
2. While manually holding the sensor on the, floor (see figure 39), perform one seating drop (no data recorded) at approximately 6,000 lbf (27 kN) followed by two drops at approximately 16,000 lbf (70 kN). Record the peaks for both drops, and refer to it as position #1.
   
   Note:  If there is any concern that a 16,000-lbf (70-kN) load will damage the test pad, use a lesser load level and scale the measured deflection linearly to 16,000 lbf (70 kN).
   
   
3. Repeat the above procedure with the deflection sensor positioned at 7.8, 11.7, 17.55, and 23.4 inches (200, 300, 450, and 600 mm) from the rear of the test pad. Keep the spacing between the sensor and the rear of the FWD load plate constant at 29.25 inches (750 mm). Name the positions sequentially (#1, #2, #3, etc.).
   
   
4. Plot the peak deflection at 16,000 lbf (70 kN) versus distance from the joint or the line. Find the distance where the deflection is roughly 20 mil (500 µm). Install the two stainless steel inserts in the floor at that location as described in the Concrete Anchor Installation section in appendix E.

Care should be taken to avoid placing the ball-joint anchor too close to a joint or the edge of the pavement. It is often the case that the edge raises before it deflects downward. This can lead to vibration and excess noise problems in the data. Based on experience, a minimum distance of 11.7 inches (300 mm) is recommended between the anchor and the edge of the pavement.

After the ball-joint anchor and the calibration stand have been installed (see figure 39), go through the “Set Trigger” and “Determine Number of Drops” routines in WinFWDCal. Adjust the position of FWD either closer to the calibration stand or further away as needed to get a successful reference calibration.

Once a suitable place for on-site calibration has been found, it can be used repeatedly for future annual FWD calibrations. To protect the two stainless steel inserts from filling up with dirt, it is wise to put an Allen^® head set screw in each anchor flush with the top of the holes.

Figure 39. Photo. Locating a place to install the ball-joint floor anchor.

ANNEX 7. REFERENCE LOAD CELL CALIBRATION PROCEDURE

INTRODUCTION

This appendix describes the new reference load cell calibration protocol in detail. It is the basis for the current version of the AASHTO R33-03 procedure.⁽¹⁰⁾ AASHTO R33-03 overrules any discrepancies with this appendix.

The reference load cell is a custom-made precision instrument that is capable of measuring loads within ±0.3 percent or better. However, such a high degree of precision can only be attained if this calibration procedure is followed exactly. It is essential that the reference load cell is calibrated using a universal testing machine that is properly maintained and accurately calibrated.

The reference load cell and its signal cable, the associated signal conditioner, and DAQ should be considered a system of instruments that should be calibrated and used together. The load cell should be calibrated to a maximum load of 24,000 lbf (100 kN).

This procedure is written with both U.S. customary and metric units shown. The calibration operator should choose one unit system and follow the procedure using the values shown. The values are not meant to be a direct conversion. The values are chosen to provide regular steps and ranges with whole numbers where possible.

This procedure has been automated and is included in the RefLCCal computer program.⁽¹⁸⁾

FREQUENCY OF CALIBRATION

Calibration of the reference load cell should be performed at least once per year. It should also be performed immediately after any of the machine screws that attach the load measurement links to the upper or lower plates are loosened. Calibration may also be necessary if the load cell fails to pass the unbalanced zero test during FWD annual calibration.

EQUIPMENT

The following list provides information on the necessary equipment for reference load cell calibration:

• Universal testing machine.
  
  A static testing machine (hydraulic or screw-powered) with a load capacity of 100,000 lbf (500 kN) or more should be used. Although the reference load cell will only be calibrated to a capacity of 100 kN (24,000 lbf), the higher capacity of the testing machine assures that the test frame will be adequately rigid. The testing machine should have several load ranges, among them a zero–30,000 lbf (zero–130 kN) range. Slightly different ranges (i.e., zero–24,000 lbf (zero–100 kN)) would be acceptable as long as the entire range of the load cell can be calibrated on a single range of the universal testing machine. Care must be taken to avoid overloading the reference load cell during calibration.
  
  Do not use a servo-controlled, closed-loop testing system for this purpose. In general, such equipment does not provide a high degree of accuracy that is required for this calibration.
  
  
• Special wood/aluminum bearing blocks for placement above and under the reference load cell.
  
  
• Custom manufactured load cell with its signal cable, associated signal conditioner, and DAQ. The load cell has three measurement links contained in aluminum housing.
  
  
• A Vishay 2310 signal conditioner, which should be the same signal conditioner that will be later used in the FWD calibration procedure.
  
  
• Keithley model KUSB-3108 DAQ, which should be the same A/D board that will be later used in the FWD calibration procedure.
  
  
• Pushbutton trigger for activating the data acquisition system.

Equipment Calibration

The universal testing machine should be calibrated annually by a certified technician according to ASTM E-74.⁽¹⁹⁾The calibrated machine should have a certified accuracy of 1 percent or better. The load indication system used for calibration should be traceable to NIST. The calibration certificate should be evaluated using a multinomial regression procedure to develop an adjustment algorithm (up to fifth order) that adjusts the indicated load on the universal testing machine to the corrected NIST traceable calibrated load. The load calculated by use of the adjustment algorithm is referred to herein as the adjusted load, while the load indicated on the testing machine dial is referred to as the indicated load.

The testing machine calibration coefficients and the date of calibration of the testing machine should be entered into the FWDCalCenterConfig.ini file used by WinFWDCal prior to calibrating a reference load cell.

The Vishay 2310 signal conditioner amplifier should be balanced according to the procedure described in the manufacturer’s instruction manual. With the signal input terminals shorted together, at gain 100, the alternating current (a/c) noise on the ±10-V output terminals should be 1 mV or less.

Equipment Preparation

Load Cell Conditioning

A new load cell or one that has had the lid removed must be conditioned before being calibrated.

1. Use a torque wrench to tighten the machine screws on the top and the bottom of the load cell to 100 lbf-inch (11.3 N-m). Apply at least 100 conditioning drops on the load cell from the 16,000-lbf (72 kN) load level with the FWD. Remove the machine screws one at a time, apply medium strength Loctite^® to the threads, and torque to precisely 100 lbf-inches (11.3 N-m).
   
   
2. Apply another 100 conditioning drops on the load cell from the 16,000-lbf (72-kN) load level with the FWD. Record the unbalanced zero for the load cell after each 25 drops. It should change less than 1 mV during 25 load cycles. Continue applying additional load cycles until the unbalanced zero stabilizes, and apply at least 20 cycles of 24,000 lbf (100 kN) to the load cell with the universal testing machine. Apply additional cycles, if necessary, until the unbalanced zero stabilizes.

Equipment Inspection and Setup

1. Inspect the reference load cell carefully before calibration. Verify that the cable and the cable connectors fit and lock tightly and that there are no breaks in the wires. Verify that the machine screws on the load cell are tight.
   
   
2. Verify that one of the wood/aluminum bearing blocks has a ribbed rubber pad cemented to it. If the edges of the pads are loose, use automobile weatherstrip cement to reattach the pad.
   
   
3. Install a spherically seated bearing block in the cross head of the universal testing machine.
   
   
4. Make the following settings on the front panel of the Vishay 2310 signal conditioner:
   
   
   • Excitation switch ON.
     
     
   • Excitation voltage set to 10 V.
     
     
   • Filter set to 1,000 Hz.
     
     
   • AC IN button fully extended (e.g., out).
     
     
   • Set gain initially to 4.2 x 100.
     
     Note: If the reference load cell has been previously calibrated, the initial gain value may be different. If so, set the Vishay gain to the most recent value.
     
     
   • Auto balance switch off.
     
     Note: Verify that the tape playback switch on the rear panel of the signal conditioner is off. Position the signal conditioner and the computer several feet apart near the testing machine and attach them to the a/c line power

5. The load cell and signal conditioner should be connected and powered for at least 30 min before performing the calibration procedure. This ensures the electronics are properly warmed up.

Calibration Procedure

Perform three calibration trials according to the following procedure. WinFWDCal must be used in conjunction with the following step-by-step procedure:

1. Hook up all cables and warm up the equipment for at least 30 min. Turn on the computer and initialize WinFWDCal. If a hydraulic universal testing machine is used, turn the pump on and allow the machine to warm up for at least 30 min.
   
   
2. Place a wood/aluminum bearing block with no rubber pad in the center of the testing machine platen.
   
   
3. Place the reference load cell on top of the bearing block with the support feet down (i.e., in contact with the top surface of the lower bearing block).
   
   
4. Place the second bearing block on top of the load cell with the cemented rubber pad down (i.e., in contact with the top surface of the load cell).
   
   
5. Carefully align the edges of the load cell and the two bearing blocks, and center the system under the spherical loading block of the testing machine.
   
   
6. Set the testing machine on a range equal to or slightly larger than 24,000 lbf (100 kN). Apply an indicated load of 24,000 lbf (100 kN) to the load cell three times. Apply the load at a rate of approximately 5,000–10,000 lbf (22–44 kN) per minute.
   
   
7. Temporarily remove the upper wood/aluminum bearing block. Set the auto balance switch on the Vishay 2310 signal conditioner to off. Read and record the unbalanced zero voltage using the pushbutton. If this voltage is in excess of ±5 V, the load cell may have been damaged by yielding, and it should be returned to the manufacturer for repair. If the load cell has been previously calibrated, verify that the difference in the unbalanced zero voltage is no more than 100 mV, or 5 percent of the previous value.
   
   
8. Briefly push down the auto balance switch on the signal conditioner to the reset position and release it to the on position. Adjust the trim knob until the KUSB board reads 0.0 V.
   
   
9. Replace and align the upper bearing block rubber pad down. Apply a load of 24,000 lb (100 kN), and check the output of the load cell is between 97 and 99 percent of full scale on the data acquisition system. In the case of the Keithley KUSB-3108, this will be between -9.7 and -9.9 V. Release the load, and record the gain setting.
   
   
10. If the voltage is not in the correct range, adjust the gain knob on the signal conditioner in 0.1 increments until the signal conditioner output is between 97 and 99 percent of full scale. (A gain of 4.20 x 100 is acceptable. A gain of 4.25 x 100 or 4.21 x 100 is not.) In no case should the output at 24,000 lb (100 kN) exceed 99 percent of full scale. The largest gain that does not produce an output above 99 percent of full scale should be used.
    
    
11. If the gain setting is changed, repeat steps 7–9.
    
    Note: When the load is released, the indicted voltage will not read exactly zero because it was zeroed before the upper bearing block was put in place. Do not re-zero the signal conditioner at this point.
    
    
12. Carefully zero the universal testing machine. Use the push-button trigger to record the reading at the zero load level.
    
    
13. Apply load at a rate of 1,000 lbf (5 kN) per minute. Use the pushbutton trigger to record the readings at 1,000 lbf (5 kN) intervals up to a maximum indicated load of 24,000 lbf (100 kN). While releasing the load, record a reading at 12,000 lbf (50 kN) and at zero pound loads.
    
    
14. Remove the upper bearing block. Use the pushbutton to record the signal conditioner calibration voltages for +B and -B shunt values on the Vishay. Set the auto balance switch to off and again record the unbalanced zero voltage. This reading should be within 10 mV of the earlier reading. If it is not, repeat the calibration procedure.

Data Analysis

WinFWDCal will perform the data analysis for each trial. It will use a stepup regression utility to calculate a fifth degree polynomial of the form as follows:

Figure 40. Equation. Load cell calibration algorithm.

Where:
Y = The adjusted load calculated from the universal testing machine indicating load including the weight of the upper bearing block.
V = The load cell voltage.
A_i = Coefficients determined by the regression.

Evaluate the results according to the following acceptance criteria:

• The standard error of the regression should be less than ±20 lbf (±0.1 kN).
• Each of the coefficients should be statistically significant.

After completion of at least three acceptable trials, WinFWDCal will pool the data for all three trials and determine regression coefficients based on the combined data, and the calibration is complete.

The final set of calibration coefficients should be evaluated according to the above two criteria. In addition, the three sets of data should be random, neither steadily increasing nor steadily decreasing. This should be verified by reviewing a plot of the residuals versus the fitted values from the regression.

Entering the Coefficients in WinFWDCal

The load cell coefficients should be entered in theFWDCalCenterConfig.ini file. Any of the coefficients that are not found to be significant should be entered as 0.0.

When the regression coefficients are entered in the FWDCalCenterConfig.ini file, the unbalanced zero, the +B and -B calibration factors, and the load cell signal conditioner gain factor should also be entered. This information is used to validate the load cell during FWD calibration.