A very simple but quite informative graphic is the actual value chart of individual values. The single values are plotted on the y-axis whereas the x-axis is a constant scale. So the x-axis is either labelled with a consecutive number of the individual values, date and time of their recording, batch number or other additional data such as machine or operator. It is important to consider the actual task of the value chart when selecting the labeling of the x-axis. In addition, it is also advisable to allocate an event to single values in order to describe the situation comprehensively. Events are e.g. tool change, change of material and other changes regarding the machine or the process.

The actual value chart of individual values provides information about typical aspects of process behaviour. It is easy to detect values outside the specification limits, trends, periodicity and sometimes even outliers.

This topic describes the special settings that only apply to this graphic, including the explanation for the value chart in solara.MP and the special value-charts in solara.MP. There are various value charts in the different products, these are described in separately.

1    Value chart

In this document, only the graphic settings specific to the value chart as well as special handling are dealt with.

In the process analysis module, the value chart exists in various sub-variants, which have fewer setting options. These are roughly explained at the end of the document, as well as the value charts in solara.MP

2    Graphics settings

2.1     Info - Window

Activated, the info window is displayed next to the value chart. When moving the mouse over a measured value, its information is displayed in the info window.

In change mode, the info window can be configured. The possible selection is reduced, here one should also limit the additional data level

A note is given on deactivating the info window: If the value chart was saved with the info window activated, the info window must be closed with the option "Info window off" to permanently deactivate it, and then the value chart must be saved. Simply closing the info window is not sufficient.

2.2     Sample display

In different variants, one can make the individual subgroups recognisable in the process analysis.

2.3     Additional moving averages

Regardless of the subgroup type, an applicable average can be displayed. The number of values or a moving formula can be set. The moving average is shown with a red line.

2.4     Body mode

The body mode was created to display characteristics measured in the negative range in the positive range.

Prerequisite

Entry 2 must be written in K2203. This defines a characteristic as "determined in negative range in body mode. The nominal size must be smaller than 0. The characteristic values are still displayed with negative measured values, but rotated.

2.5     Optimise defocus

The setting of the defocus should only be made in projects on the instruction of the project contact person.

2.6     Draw symbols

With the option the symbols of the measuring points can be hidden

2.7     Display of invalid values

Invalid values are measured values which have been deactivated, e.g. by plausibility limits. Such invalid measured values are shown in the standard with dashed connecting lines:

Another option is to hide the value with a solid connecting line:

Or to show this as an interruption in the value chart

2.8     Allocation for additional data

With the allocation for additional data, the characteristics of selected additional data can be graphically displayed in the value chart. The example shows a breakdown by batch number.

If the measured values have alternating additional data, where such a superimposed display can easily become confusing, an individual characteristic can be displayed alone by clicking on the label in the legend. In this example, the gage "Hardness tester" is to be considered; for this purpose, it is clicked on in the legend:

After that, all other characteristics are deactivated for a closer look at the selected characteristic.

With another click in the legend, all expressions are displayed again.

With the two options "Activate / deactivate division" a previously defined and saved division can be called up quickly:

In the main option, the dialogue for activating and selecting the additional data field is opened.

A more precise description is needed for the breakdowns by "period" and "recurring period". To show the difference, the same time unit "day" is used, even if it makes less sense in the first example.

2.8.1     Allocation by period

Using the example of a data set with 3500 measured values from 4 months. An allocation by period "day" is selected. Each day is displayed as a separate block with an entry in the legend:

2.8.2     Allocation by recurring period

As a counter-example, the "weekday" is chosen for the recurring time periods. In this case, all 7 days of the week are shown in the legend.

The allocation for additional data is a purely graphic allocation in the value chart and has no influence on the calculation. If the various additional data characteristics or time ranges require separate evaluations, this can be done when reading from database with the option "Automatic division".

2.8.3     Division via button bar

A predefined button bar is available for quick use of the graphical division (only for additional data, not for time periods), which can be integrated in the assistant, for example. This can be used for quick allocations for additional data:

2.9     Additional data change

Another possibility is the separator when changing additional data. Here, the batch number is used as an example. If there is a change of additional data, this is indicated by a blue separator in the value chart.

This option only makes sense for additional data that have a more "stable" character, i.e. that do not change constantly. Such as the classic batch number or order no.. Since no legend is used here, the additional data field could still be displayed with the axis labelling.

2.10  Quantiles

The options to display the quantiles graphically more clearly are repetitions of the settings under "Background".

3    Sub-types of the value chart

In the modules of sample analysis and process analysis, there are various subtypes of the value chart (depending on the module), which will be briefly discussed here

3.1     Value chart (section)

The "Value chart (section)" has the number of measured values as an additional option, here as an example: The last 10 measured values.

This value chart is mostly used in overview graphics, as the classic value chart can become confusing with a larger number of measured values, especially if the focus is to be on the new data of the current production:

3.2     Value charts for statistics of the individual subgroups

If required, value charts for the characteristic values of the individual subgroups are available in the process analysis (value chart for average values / minimum values / maximum values...).

These value charts are in no way to be seen as a substitute for the quality control chart!

3.3     Trajectories for attributive characteristics and their sample data

Especially for attributive characteristics, value charts of the sample size, the errors per unit and the errors in % are available:

3.4     3D value chart

In the 3D value chart, the individual subgroups are displayed starting with the first value of the subgroup "backwards".

4    Value charts in solara.MP

Various value charts are available in the solara.MP programme for visual assessment. The following pages explain the interpretation of these value charts as well as new configurations.

4.1     Value chart procedure 1

Data set: Test_14.dfq

The value chart of method 1 shows, in addition to the reference measurements, the characteristic values x _g(measured average), x _m(actual value of the standard) and the reference value of the tolerance. (x _m+-0.1*T).

4.1.1     Bias

From the graph, the systematic measurement deviation can be read, because it is the difference between x _gand x _m.

4.1.2     Reference

The reference is given as half of the reference share given in the strategy.

With 0.2 * reference value therefore +/-0.1

At 0.15 * reference value +/-0.075

If the process Variation is selected as the reference variable, the labelling changes

4.2     Value chart procedure 2

Data set: Test_16.dfq.dfq

The value chart for method 2 shows the measurements of the individual operators in various sections. Over the entire value chart, it can be recognised whether the part variation is sufficient (important, among other things, for the calculation according to MSA 3rd Edition).

4.2.1     Display titel

If the titel is activated ("Image title on"), the designation is displayed above each section.

4.2.2     Measured values

The repeat measurements of an operator on a component are displayed one after the other.

4.2.3     Graphics options.

Newly added to what is possible in qs-STAT are the inspector details in procedure 2:

Beforehand, the warnings mentioned in the document "FAQ_General-Settings" should be repeated:

Data protection and the burden of proof In most European countries it is NOT allowed in classical industry to carry the operator for the measured value in Measurement system analysis. Keeping operator information side by side could be used to judge the "performance" of different operators against each other. This is often seen as a violation of applicable law for the protection of employees. However, this is countered by the obligation to provide evidence. During audits, the responsible persons are often requested to provide written evidence of the proper performance of a Measurement system analysis. It must therefore be proven that the Measurement system analysis has been carried out over time by different operators. It is not necessary to keep auditor information on the measured value. The information can be carried along, for example, by a general formulation of the inspector information in the characteristic field "Remarks". The information in the characteristics mask can then be printed on the reports.

If the display of the operator name is permitted in the general system settings, this can be activated via the options here.

4.3     Value chart procedure 3

Test example Test_17.dfq

Since no different operators carry out the measurements in method 3, only the pure repeat measurements are shown. Over the entire value chart, it can be recognised whether the part variation is sufficient (important, among other things, for the calculation according to MSA 3rd Edition).

4.3.1     Measured values

The repeat measurements of a component are displayed one after the other

4.4     Value chart Attributive

Data set: MSA3_ATTR.dfq. Subfolder MSA 3 Edition

For attributive data sets, a complete value chart is only displayed if reference values are available.

4.4.1     Consistent refusals

Values that all operators have consistently rejected in all measurements, consistent with the statement of the reference value, are marked with a red square.

4.4.2     Matching assumptions

Values that all operators have accepted consistently in all measurements, consistent with the statement of the reference value, are marked with a green arrow.

4.4.3     Non-conformities

Values for which the operators have made no concordant statement or a statement contrary to the reference value are marked with a cross. The range in which these values lie is the grey area.

4.4.4     Grey area

The range that the grey area occupies as a percentage of the tolerance is output as R&R (GRR).

In the chart "Individuals, ascending order" this grey area can be seen again

4.5     Value chart stability

Data set GC_STAB.dfq

The value chart for stability corresponds to that of the process analysis.

Due to the analysis of the stabilities, the value chart is of little importance here; the QCC should be considered here.

4.6     Value chart linearity

Data set GC LIN.dfq

The measured values for each reference are shown in the value chart of the linearities.

5    Special value charts in solara.MP

Various value charts are available in solara.MP for visual assessment. The interpretation of these value charts is explained below. The basic functions correspond to those of the normal value chart

5.1     Value chart averages

The value chart of the averages shows a section for each operator. (Operator A: values 1A to 10A; Operator B: values 1B to 10B...). Each individual section shows the averages of the measurement of the components of each operator.

5.2     Value chart deviations

The value chart of the deviations shows a section for each operator. (Operator A: values 1A to 10A; Operator B: values 1B to 10B...). In each section, "number of measurements + 1" lines are shown. These are the individuals of the operator and the average calculated from them (4 lines for 3 measurements). The lines of the measurements are numbered so that the respective measurement of the operator can be recognised.

Each line shows the deviation of the individual measured value from the average of this measured component of all operators, all measurements. The values mask shows the difference between the value outlined in red and the average of all values outlined in blue.

5.2.1     Mathematically explained:

The first (blue) measurement series of operator A is formed:

The second (red) measurement series of operator A is formed:

The third (light blue) measurement series of operator A is forming:

The green, unnumbered average line of operator A is formed:

5.2.2     Graphic interpretation:

Due to this normalised representation, which does not take into account the part Variation, the calculated characteristic values AV and EV, or rather the problems with this measurement system, can be recognised graphically in this graphic.

In contrast to operator A and C, the variation of operator B is too great:

Operator C, however, has a downwardly shifted location compared to operators A and B:

5.3     Value chart variations

The value chart of the variations shows a section for each measurement run (measurement 1: values 1A, 1B and 1C; measurement 2: values 2A, 2B and 2C...). In each section, one line is shown per operator. These are the individuals of the operators.

The mathematical explanation follows that of the graph "Value chart deviations", however, in this graph the relation to the measured components can be better recognised.

Also in this graph it can be seen that operator C (light blue) has a downward shift in position in contrast to operators A and B and that operator B (red) has a larger variation than operators A and C in most measurements.

5.4     Overview value charts

For many single-characteristic graphs, 2 overviews are available especially for the procedure. These are explained on the basis of the value chart

5.4.1     Clarification of the use of language

What does "part centre" mean ?

The part mean is the average of a measured part:

In the values mask: the average of each individual measurement series

In the "Value chart overview related to part centre", these 9 values are the deviations of the measurements related to the average of all measurements. The average is calculated over all values, including the value under consideration.

What does "reference" mean ?

In almost all classical methods 2 the graph shows no content:

In a special procedure 2 with the specification of reference values (determined on a higher-quality measuring device, as reference components), a linearity can be output within the procedure 2. In order to obtain a meaningful evaluation, it is absolutely necessary to use components that cover the entire range to be considered (in the classical sense: the tolerance) and possibly even exceed it.

Necessary settings in the strategy:

In the area of data recording, reference measurements must be allowed. Repeat measurements of the references are also possible:

With this, the graphic evaluation would be nicely possible.

If this is also to be assessed mathematically, it is recommended to activate the GMPT requirements for linearity and the minimum coefficient of determination in a sub-strategy.

These claims should not be activated in the main strategy, as otherwise reference values are required. If the procedure is not known in detail, this must be learned in a training course.

The number of reference measurements can be specified on the characteristics mask:

These are then entered BEFORE the operator measurements, similar to the linearity data sets.

The graphic "Value chart related to reference" is then also filled graphically.

5.4.2     Miscellaneous graphics

Basically, all blocks are divided as follows:

• Overview chart of all operators
• Graphic per operator
• Overview graphic of all operators and additional graphic per operator

The further separation of the different blocks:

• Graphic related to reference or part centre (first the reference is searched, if not available the part centre is used)
• Graphic related to part centre
• Graphic related to reference

Thus, the first of the overview graphs represents all variants:

6    Tips for handling the value chart

This appendix is intended to briefly address frequently asked questions about the value chart without going too deeply into "detailed training".

6.1     Hardly visible values due to outliers

In some cases there are extreme outliers. The graph itself tries to show all values and characteristic curves that are to be represented. However, as in the upper case, this means that measured values or specification limits are hardly recognisable.

Possible remedies:

Display limit:

A graphical possibility is to limit the display of the value chart to X% in addition to the tolerance:

Outlier distance

By specifying plausibility limits to remove illogical / implausible values, by manual outlier removal or by mathematical outlier search of the strategy, such values can be removed:

6.2     Starting the "Part measurement" in the parts protocol".

After identifying a conspicuous value in the value chart, there is often the wish to see the "part measurement" of this value. In other words, all values of the other characteristics of the common measurement.

For this purpose, the parts protocol can be displayed next to the value chart. A measured value can be held with the mouse and dragged onto the parts protocol. Its view then changes to this measurement.

Topic in PFD format

This topic is also available in PDF format.

English:PDF

German:PDF