|
These test results were collected on a random customer unit, S/N 1787 (Novaspense 16-4). Our customer, who requested confidentiality, contracted engineering services for a detailed statistical evaluation of their pump. The specified target dispense was 8.0 ml with a specified accuracy limit of +/- 0.5ml. For our analysis, we elected to reduce the tolerance to +/- 0.25ml. Please note, we reduced/refined the accuracy resolution by 50% for testing purposes.
The chart reflects the distribution of volumes taken and measured during that study. The definition of data points is as follows:
Model 16-4 End User (with Novaspense)
Mean-
USL-
LSL-
Cpk-
Cp-
s-
Customer/Application
Automation, system integration
Testing Definition: (What are Cp and Cpk?)
Two critical elements of test data are Cp and Cpk. Cp is a comparison of the
process variation to the part tolerance. If a process is “in control”,
the Cp will be equal to or greater than 1. It is generally recognized that
process engineers attempt to achieve a Cp of about 1.67. Cpk is a measurement
of centeredness. In a prefect situation the distribution of points is equal
on both sides of the mean and mean is close to the target. Some processes
are under control with a Cp of 2, but may have a Cpk of 0.5. This means the
entire curve will be shifted towards the LSL. Although the process in control,
a sizeable amount of scrap would be created. As with Cp, a Cpk value of 1
means the process is in control, but process engineers strive to achieve 1.67
or better.
Test Conclusions:
The Novaspense unit performed substantially beyond the specified accuracy
limits even when the requested specification limit was reduced by 50%. At
a tolerance reduced to +/- 0.25ml, the unit still achieved a Cp of 6.15 and
a Cpk of 5.7.
To the Process Engineer this is very significant. By reducing the tolerance,
you are able to reduce the use of raw material and scrap. Therefore the overall
cost of the process is reduced.
