Test Types

ProCalV5 allows for the addition of different “Test Types” to the database that may be applied to any item master record whose required calibration does not fit perfectly into the typical ProCal template.  Test types may be used to provide specific testing in addition to or instead of a standard calibration with target set points and as found/as left readings.  Examples of some calibration test types that ProCalV5 normally could not accommodate efficiently are extreme values, linearity and standard deviation tests.

 

 

Set Up Test Type Packages Screen

The Set Up Test Type Packages screen (shown in Figure 1) is used to add, modify and if necessary delete test type packages from the database.  This screen is invoked by selecting the “Set Up Test Type Packages…” option on the File menu when the System Options and Defaults screen is open.

 

 

Figure 1


Adding New Test Types

To add a new test type to your database you must first copy the required library file “Linearity1.bpl” into the folder labeled “bin” under the ProCalV5 directory (typically “C:\Program Files\Prime Technologies\ProCal V5\bin”) in Windows Explorer.  The “Linearity1.bpl” file contains all the information necessary for the test type’s automatic and proper setup in ProCal.  Once the file has been copied into the correct folder you may launch the Set Up Test Type Packages screen.  Once the screen is open you may click the Add New Record button or select “New Record” from the screen’s File menu.  This will invoke a browser screen titled Choose New Package File that automatically directs you to the “bin” folder in which all the “.bpl” files reside. Select the required file “Linearity1.bpl” and click the Open button or double-click the file name.

 

Once the correct file has been selected you will be returned to the new test type record which will already be set up for you.  In this instance you can save the record, making it available for application to all types of item master records in the system (i.e. instruments, loops, systems, test instruments, in addition to any PM records that may require a calibration).  Once a test type has been saved to the database, as long as it is marked as “Active”, it will be available for selection on the Test Type field’s drop-down list on the Calibration Points tab in all screens where that tab exists.  

 

Although you will have no ability to change the test type’s calibration requirements, you do have the ability to make some customizations to it in terms of field label changes and also giving users the ability to edit target set points and such when entering calibrations using that test type.

 

 

 

Distributing Test Types

Test Types only need to be installed in the System Option and Defaults screen once, but for all Client, Administrator, or Mobile Workstations, the “Linearity1.bpl” file must be present in the “bin” directory one level down from the ProCalV5 installation directory.

 

 

 

Definitions for the Scales and Balances Test Type Package

R – correlation coefficient 

R Squared or R2 – coefficient of determination


 

Entering Item Records Using the Linearity Test Type Package

When creating a new master record that uses the Linearity Test Type, the test type must be applied to the record through the Calibration Points tab on the master record screen.  Once a new group has been added to the record, click the Test Type drop-down list and select “Linearity” from the available test types.

 

 

Figure 2

 

  

Once the test type has been selected, the Test Point Group inputs will change according to the settings defined by the Linearity TTP library file, as shown in Figure 2:

 

 

 

This test type package will calculate a Linearity “correlation coefficient” from the As Found Readings taken for a set of test point inputs.   The calculated value is then used directly as the R output value, and squared to produce the R Squared value.  The pass/fail can be done against R (the correlation coefficient) or R2 (coefficient of determination).  If the absolute value of the specified value to test against correlation coefficient is less than the minimum, the test fails.

 


 

The Linearity test is configured by specifying the Input Type, Input Resolution, Ranges, Output Type and Output Resolution.  The number of Cal Points MUST be entered, for it corresponds to the number of readings used for the Linearity calculation.  The remaining standard Test Point Group Input fields will be removed when the Linearity TTP is selected, 

 

 

Figure 3

 

 

The Additional Test Type Inputs used in this test can be entered into the corresponding fields in the Additional Test Type Inputs grid of the Group Options box, accessed by clicking the Show Group Options button on the right side of the screen, as shown in Figure 2.  The Additional Test Type Inputs grid interface is shown in Figure 4.

 

 

Figure 4

 

The Additional Test Type Inputs used in this test are as follows:

  • Minimum Accepted Value – Minimum value of a passing correlation coefficient or coefficient of determination.  A value of 1.0 represents a perfect linear correlation between the As Found values recorded during the execution of the test and the Actual Input values that specify the straight line being tested against.  

NOTE: A device that has inherently high accuracy with respect to Linearity, such as an Analytical or a Micro Balance, will generate an R and R2 so close to 1.0 that this test may not provide usable data.  In this case a standard accuracy test can be specified with a ProCalV5 Test Point Group without the Linearity TTP utilizing a Linear Correlation between the As Found values and the Actual Input values that specify the straight line.

  • Resolution – List of available Resolutions (1 to 0.0000001) to select from for the Resolution of the calculated correlation coefficient and coefficient of determination.
  • Value to Check – Select either R (correlation coefficient) or R squared or R2 (coefficient of determination) as the value that ProCalV5 will compare to the Minimum Accepted Value in order to make a Pass/Fail determination for the test.

 


 

Entering Calibration Records

Once a Test Point Group has been created for a master item record that uses the Linearity Test Type, a calibration record can be created and data can be entered on the Test Results tab on the Calibration screen.  It should be noted that the required inputs must first be entered on the master item record as detailed above.  The Test Results tab for a Device Master Record utilizing the Linearity TTP is shown in figure 5:

 

 

Figure 5

 

After making any necessary adjustments to the Actual Inputs, In Val entries to correspond to the actual Inputs used by clicking the Modify Specs button (see the ProCalV5 Main Application User Manual for a discussion of this functionality), the Technician will enter data for each Test Point in the Readings, Fnd As fields.  When all As Found data has been recorded or when ready, the Technician will click the Recalculate Test Type Outputs button and the Linearity TTP will display the calculated R and R2 values in the Test Type Outputs box as shown in figure 5 above.  

 

The Linearity TTP will also display the Pass/Fail determination based upon the As Found values in the Test Type Outputs box by scrolling down in the box, as shown in figure 5a below.   Complete the Calibration Record as you would any standard ProCalV5 Calibration Record according to the Work Flow practices of your Organization.

 

 

Figure 5a

 

The exact specifications for this Test Point Group can be viewed by pressing the Show Group Details button, opening the Test Point Group Details window (Figure 6 below). The fields on this window are modifiable only on the device Master Record.

 

 

Figure 6


 

Calculated Outputs

 

These values are generated by the ProCalV5 Linearity TTP using the As Found data collected on the Calibration screen.

 

  • R – Calculated Correlation Coefficient
  • R Squared – Calculated Correlation Coefficient Squared (R2)
  • Result – “Pass” or “Fail” determination from the As Found data

 

 

Calculations

 

All calculations are performed using As Found readings:

  • Correlation Coefficient – given a set of Input Signals I, a set of Readings O, the average of the Input Signals Iavg, the average of the Readings Oavg, and the number of test points m, the correlation coefficient is determined as follows:

                                     ∑ n = 1 to m of (In – Iavg)(On – Oavg)                    

R = 

√( (∑ n=1 to m of (In – Iavg) 2) * (∑ n=1 to m of (On – Oavg) 2))

  • Results – are calculated by finding both the R and R Squared values and comparing the selected one to the Minimum Accepted Value.

 

 

Pass/Fail Determination

 

As Found Failure (Found out of Spec) – the “Found out of Spec” Calibration Pass/Fail logical condition is not determined and therefore not reported by the Linearity TTP.  The As Found test will by default always pass.

As Left Failure (Cal Failure) – the “Cal Failure” Calibration Pass/Fail logical condition is determined and reported by the Linearity TTP based on the As Found data collected as follows:

  • When the “Value to Check” input equals R and the R value is less than the “Minimum Accepted Value”, the Linearity test fails.
  • When the “Value to Check” input equals R Squared and the R Squared value is less than the “Minimum Accepted Value”, the Linearity test fails.