SPC Charts for Variables Data

As a true preventive tool, control charts for variable data provide the measure of process improvement. Since all applications are not the same, Quality America is proud to offer a wide variety of these tools to meet your specific needs. From short run to continuous flow, administration to maintenance, your profitability can be improved.

SPC-PC IV offers the following charts for variable data:

• X Bar/Range The classic Shewhart control chart, with provisions to:
  • Allow missing or deleted data, adjusting control limits accordingly.
  • Permit control and/or warning limits at the user-specified sigma level.
  • Automatically exclude out of control points from calculations, at the user’s option.
  • Include the effect of Western Electric, Nelson, and user-defined Run Tests.
  • Perform Short Run Analysis.
  • Set multiple ranges for control, so process changes can be controlled to their new level.

• X Bar/Sigma With the same flexibility found in the X Bar/R charts, these charts provide a more precise indicator of standard deviation.

• Individual X/Moving Range Preferred by many customers for their easy interpretation, decreased sampling costs, or when subgroups do not provide a real measure of process variation, SPC-PC IV provides the same flexibility found in the X-Bar charts, plus:
  • Your choice of control limits, using either the Normal distribution or Quality America’s unique Johnson Control Charts for non-normal processes.
  • Control limits and Run Tests based on the selected distribution.

• Process Capability Analysis To truly measure the capability of your process both your process output and your requirements must be correctly stated. SPC-PC IV provides analysis options capable of meeting your requirements. SPC-PC IV uses any of the following distributions, at your option:
  • Normal.
  • Johnson: provides one of the Johnson family of bounded or unbounded curves fitted to the data (includes the normal and log normal).
  • Folded Normal: typically used for TIR measurements, such as concentricity, roundness, flatness, etc.
  • Rayleigh: used in ANSI Y14.5 positional measurement systems.
  • Weibull: used extensively to model reliability and particle size distribution.
  • Calculations include Cp, Cpk, Cr, Cpm, and Z values, as well as predicted yields. Confidence limits are available for capability indices. SPC-PC IV also provides comparison of non-normal and normal assumptions.

• Short Run Analysis Applying standard control limit constants to a short run of only fifteen subgroups of size five will double the probability of “false alarms” and result in tampering with an in-control process. Short Run uses multiple parts (or doctors, billing types, etc.), each constituting a distinct “run,” to determine common characteristics of the process for all runs. Available as an option in X-bar, Individual-X, EWMA, Process Capability, P, Np, C and U charts.

• Scatter Diagram To examine the relationship between two different characteristics, regression techniques are used to estimate linear models and the correlation between the characteristics. Confidence Limits may be used to identify data which does not fit the regression model.

• CuSum When the cost of process shifts or sampling is high, many customers use the CuSum chart for increased sensitivity to small process shifts, or for comparable protection at lower sampling costs.

• Autocorrelation Function An analysis tool rather than a control tool, the Autocorrelation function provides a means of determining the extent to which current process conditions are dependent on previous conditions. If autocorrelation is significant, standard control charting may not be used since the independence assumption is violated. In these cases, the EWMA chart for drifting means may be used, or the autocorrelation parameters may be used with QA, Inc.’s spreadsheet functions to model the process. Errors from the model may then be controlled using standard control charts.

• Exponentially Weighted Moving Average Similar to the CuSum in detecting small shifts in an otherwise stable process mean, the EWMA chart provided by QA, Inc. may also be used to control processes with a slowly drifting mean using Montgomery’s technique for autocorrelated processes.

• Moving Average/Moving Range or Moving Sigma These tools are useful for controlling processes with inconsequential cyclical patterns which would otherwise produce false warnings on standard charts.

• Multivariate Controlling several related characteristics individually will not always signal true shifts in the process. SPC-PC IV’s Multivariate Analysis provides the T2 and SPE (Squared Prediction Error) Control Chart for detecting process shifts, and Contribution Charts for identifying the key process variables responsible for the shift.