APPENDIX F: DEFINING THE OPTIMUM PROCESS
CD-A DUAL PULSE RESISTANCE WELDING POWER SUPPLY
990-430
F-11
Criteria for Success
Before running the series of experiments, the user must establish an acceptable window for energy, time,
and force, thus preventing voided results. It is common practice to include one or all of the above variables
in a Screening DOE. This is only recommended if sufficient understanding has been established for the
other application and process variables that can impact quality Users should first try to screen out all
common application and process variables that require further exploration from the results of the “look
see” mini experiments and then include the three key welding variables (energy, force and time). Several
Screening DOE’s may be required.
Results should be interpreted carefully. Typically, one would look for the highest result in terms of quality
with the least variation. A Screening DOE provides only a measurement that indicates the relative
importance of a parameter and not the ideal setting. Factorial DOE’s should be used to establish the
correct or best setting for a parameter once many of the other variables have been screened and fixed.
This is also the time to assess the measurement accuracy and consistency of the test method and procedure.
Variation in test method can invalidate the test and lead to misinterpretation of results.
What are Factorial DOE’s?
The purpose of a Factorial DOE is to narrow in on the optimal setting for a particular parameter. This
method is generally used when the critical or main key variables have been identified, and we need to
establish the best settings for the process. A factorial DOE may also give an indication as to how wide
the acceptable weld window is in relation to quality requirements. We recommend data be gathered from
a monitoring perspective so that this can provide a starting point for establishing a relationship between
quality and the monitored measurement parameter.
Criteria for Success
Critical parameters should be identified from the list of unfixed variables left from the Screening DOE’s.
A mini-experiment may be required establishing reasonable bounds for the combination of parameters to
be tested. This will prevent void data and wasted time. At this stage, it is useful to record multiple relevant
quality measurement or inspection criteria so that a balanced decision can be reached. For example, if
part marking and pull strength are the relevant criteria, a compromise in ideal setting may be required.
As with all experiments, the test method should be carefully assessed as a potential source of variation
and inconsistency. Once the optimum parameters have been established in this series of experiments, a
validation study can be run which looks at the consistency of results over time. It is good practice to build
in variables such as electrode changes and cleaning, as well as equipment set up by different personnel.
This will ensure that the solution is one that can run in a real production environment. Welded assemblies
should be tested over time and under real use conditions to ensure that all functional criteria will be met.
Validation testing is usually required to prove the robustness of the process under production conditions.
