How small an increment can the measurement system measure? Perform a discrimination study.

Check one system per system family.

Preferred rule for acceptance:

> 5 increments between zero and the UCL on the production/service process range control chart.

Secondary rule:

> 10 increments across the 6s width of the product/service variation curve.

How wide is the measurement system error curve? Perform a gage R&R study.

Check a representative sample per system family. A minimum of three systems with a maximum of 25% to 30% of the total.

Preferred rule for acceptance:

PV/TV ratio > 95%

Secondary rule:

% RR/Tol. < 10% good

% RR/Tol. < 30% but > 10% gray area

% RR/Tol. > 30% bad

How accurate is the measurement system (on average, does it provide the "true" value)? Perform a bias (accuracy) study.

Check a representative sample per system family. A minimum of three systems with a maximum of 25% to 30% of the total.

Preferred rule for acceptance:

% Bias/Tol. < 5% for noncritical and < 2.5% for critical measurement systems

Secondary rule:

% Bias R&R < 30%

Is the measurement system consistent over the operating range? Perform a linearity study.

Check a representative sample per system family. A minimum of three systems with a maximum of 25% to 30% of the total.

Preferred rule for acceptance:

Absolute value of % linearity < 5%

Is the measurement system consistent over time? Create a control chart.

Check all systems within a system family.

Preferred rule for acceptance: Measurement system remains in control. If measurement system goes out of control, investigate, improve the system and re-perform short-term assessment studies (precision, bias, and linearity).