To produce products of high quality, changes are often needed to the manufacturing process. These refinements could be to material quality or to process conditions.
To produce products of high quality, changes are often needed to the manufacturing process. These refinements could be to material quality or to process conditions.
The Taguchi statistical analysis is one means of providing an analytical solution to the calculation of the effects of process variables.
The National Physical laboratory has managed a programme on developing materials measurement methods for adhesives in which the Taguchi method has been applied to aid companies in the manufacturing process. This programme was sponsored by the Department of Trade and Industry. Analysis was conducted by Xyratex (formerly IBM Havant).
The Taguchi statistical analysis focused on the adhesive bonding process used in the assembly of an actuator for a computer hard disk drive. The actuator is a linear type where high quality bearings on a carriage assembly move along ultra smooth rails in order to accurately position the read/write heads over the disks. The heart of this actuator is a core/rail assembly in which fine zirconia rails are bonded to a nickel plated iron core using a single part anaerobic adhesive. During initial product development, this bond failed approximately 50% of the time during further assembly processes in the manufacturing line.
The manufacturing process resulted in a variable bond strength frequently falling below quality targets. This meant that high inspection levels were needed and high rejection rates were suffered.
The objective of the analysis was to create a bond strength better than 2kN. Properly designed experiments assessing each variable at typically two levels, enable the array to separate the effect of each variable within a matrix of manageable size. Thus it offers the opportunity to identify and rank the process variables which are used to define the process conditions, and to identify the individual variables and combinations which can lead to optimum process performance. Formulas are used to convert the process variables into a final solution of their effects. Seven process variables were identified and the associated levels determined as given below.
|
Experiment |
Factor |
Level 1 |
Level 2 |
|
A |
Parts temperature |
20oC |
50oC |
|
Time in fixture before cure |
5min |
15min | |
|
C |
Time to cure |
30min |
120min |
|
D |
Time in solvent wash (cleans excess adhesive) |
5min |
20min |
|
E |
Time in oven (final cure) |
30min |
60min |
|
F |
Oven Temperature |
20oC |
100oC |
|
G |
Time post oven |
4 hours |
12 hours |
Fine Zirconia rails
Nickel plated iron core
Single part anaerobic adhesive
Not specified. Control parts temperature
Apply adhesive. Initial air cure. Solvent clean to remove excess adhesive. Heat cure.
Air and heat curing
Initial air cure 30 to 120 minutes
Solvent wash 5 to 20 minutes
Final oven cure 30 to 60 minutes
Post oven 4h to 12h
The bonding was carried out in the laboratory
Analysis showed factors to be ranked in the following order, with best levels as indicated.
|
Variable Significance |
Variable |
Setting |
|
1 |
Oven temperature |
100°C |
|
2 |
Time to cure |
120 min |
|
3 |
Time in solvent |
5 minutes |
|
4 |
Time in oven |
60 minutes |
|
5 |
Time post oven |
4 hours |
|
6 |
Time in fixture |
15 minutes |
|
7 |
Parts temperature |
20°C |
A confirmation trial gave a mean bond strength of 9.3kN against a specification of 2kN minimum. The process reliability was determined to be excellent with the chance of falling below specification to be less than 1 in 109.
The Taguchi analysis allowed the examination of production variables on core/rail bond strength. It also defined a cleaning stage in the adhesive bonding process which effectively removed excess adhesive from the rail surfaces which contact the carriage bearings.
The trial was judged to be a complete success, and delivered a result in much less time than alternative approaches.
Arguably the process is simpler and better controlled than many packaging applications because of the quality of materials and the time available. However it clearly illustrates the value of the method. Taguchi has proved to be a useful tool in optimising an adhesive bonding process. In the case of this core/rail bond, more than 350,000 units (700,000 bonds) have been produced with no failures.
DTI MTS programme adhesives – Dissemination Guides ADH5CS1A.DOC
NPL / ESR Technology Limited
Analysis was conducted by Xyratex (formerly IBM Havant).
Hard disc image courtesy Seagate Limited
