Process changes in a manufacturing facility are sometimes made without systematic testing or documentation, meaning critical process variables are adjusted according to intuition, not engineering science. When customers began voicing concern to us about the product life and inconsistent quality of our Hot Ink rollers, we began approaching problems in a more orderly manner. The first corrective tool we implemented was design of experiments (DOE) software. Since that time two years ago, DOE has not only dramatically improved our Hot Ink roller consistency and impression count, it now sends a message to customers that because we listen -- and because we have a superb manufacturing tool in DOE -- consumers have greater confidence in our entire product line. The result at Coding Products, an ITW Company (Kalkaska, Mich.), is more sales and more customers who are delighted with our company.
Coding Products’ top-selling Hot Ink Rollers, pre-inked foam rollers measuring approximately one inch in diameter, are impregnated with ink before being sent to our wide variety of customers. Food manufacturers, for example, use our rollers to print expiration dates onto candy bar and other food wrappers. End users heat the rollers, releasing ink that is available in an array of colors. This application method, which has been in use for decades, alleviates FDA and environmental wet-ink printing process issues.
In our pre-DOE days, a customer might get 65,000 impressions from one of our rollers, while the next roller would produce anywhere from 30,000 to 120,000 impressions. Consistency in the number of impressions is crucial because it allows customers to schedule roller changes. Not knowing when an ink roller would run dry caused printing problems or waste by discarding usable rollers. This important customer satisfaction issue could also affect how they viewed the reliability of our other products.
DOE BenefitsTo discover the road to greater impression consistency and better customer perceptions, we selected Design-Expert® DOE software (Stat-Ease, Inc., Minneapolis, Minn., 1-800-801-7191, www.statease.com). DOE systematically provides information about factor interactions and the way a total system works. It shows how interconnected factors respond over a wide range of values, or levels, without requiring the testing of all possible values directly. Collectively, these equations, which were too complex for the nonstatistician before the advent of the PC, serve as models that predict what will happen for any given combination of values. Optimizing critical responses and finding the best combination of values is essential to successful product design.
DOE enabled us to decrease the variability of the number of impressions, something almost unattainable through our previous method of testing one-factor-at-a-time (OFAT) while holding other factors constant. To regain our competitiveness, we could no longer afford OFAT experimentation because it does not provide statistically precise results. Compared with DOE, traditional OFAT testing methods are extremely time-consuming -- and results are not always reliable.
Our DOE Goals:
- Achieve consistent ink impressions from roller to roller
- Increase average number of impressions
- Decrease processing time
- Standardize run conditions
- Develop tests to determine roller endurance.
We began by testing the average number of impressions inked by 25 rollers that were pulled randomly from our current manufacturing conditions. The average was 65,500 impressions per roller, with an unenviable variation of 71%.
The Critical Factors:Using our product knowledge and experience, we chose five processing factors to vary. A key element of DOE is its capability to glean a great deal of factor information by directly testing only two levels from each factor – the high and low levels. The contrast between levels provides the driving force that uncovers the most dominant effects. This technique produces a powerful parallel-testing scheme that evaluates a setting’s effects. By restricting each factor’s tests to high and low levels, we minimized the number of experiments needed – resulting in significant cost savings. Within one month, we recovered the cost of our DOE software by implementing the recommended changes in our hot ink rollers.
To test the process thoroughly, we conducted 16 experimental runs. Analysis of the results revealed two proprietary factors most critical to roller performance. We optimized the process, resulting in an increase the number of impressions and decreased variability.
The ResultsDOE software disproved some original processing theories, while revealing which factors were most critical. The results were displayed in a simple format that was easy to use and understand (see Figure 1.)
The quantity of hot ink roller impressions increased by 10% and variability decreased from 71% to 26%, providing more value for our customers. Additionally, we significantly decreased our manufacturing time, thereby improving productivity and increasing capacity. During the DOE, we developed an accelerated test method for evaluating the number of impressions we could expect from a given hot ink roller, allowing us to test future improvements. Using this test method to evaluate a formulation change has already resulted in an overall 33% increase in hot ink roller impressions.
Because we now know which variables are controllable as well as the effects of their ranges, we’re focusing our expenditures on the most beneficial features. For example, we spent our allotted $15,000 improvement budget on controlling those factors that have the most affect on the positive aspects of the process. Additional, we set the less critical factors to the most cost-effective levels.
Because of the confidence we have attained using DOE, we are currently evaluating other areas of the hot ink roller product design as well as other product lines, including hot stamp foils and thermal transfer ribbons. We are also using DOE software to investigate ink formulations.
Proud HistoryIn the mid-1980’s, Design-Expert DOE software reintroduced DOE to researchers and engineers. Invented in the 1920’s, it was not widely used because of its complex hand calculations. Until the number-crunching capabilities of PC’s and the availability of reasonably priced DOE software, DOE languished in mathematics departments. Today, however, DOE software easily sets up and analyzes statistically sound DOEs in just hours – dramatically faster than the months spent testing only a handful of factors using the traditional method. DOE has allowed us to regain our competitiveness and increase our worth to customers by providing them with high-quality hot ink rollers that last one-third longer.
Contact Information:Design-Expert v5 softwareStat-Ease, Inc.2021 East Hennepin Avenue, Suite 191 Minneapolis, MN 55413 Phone: 800-801-7131, 612-378-9449 Fax: 612-378-2152 e-mail:email@example.comWeb site: www.statease.com
Coding Products, An ITW Company Jennifer Borkovich, Development Chemist e-mail: firstname.lastname@example.org
Lisa Surowitz, Process Engineer
111 West Park Dr.
Kalkaska, MI 49646
Phone: (616) 258-5521
Fax: (616) 258-6120