Introduction
Nowadays, businesses all over the world are facing the challenge of quality. With globalization, competition and demand are increasing. Quality is no longer a purely technical issue but has become a top strategic issue related to the survival of businesses.
Kyoshin Vietnam (KSVN) is a company specializing in manufacturing components for the automotive industry, with many different types of products. Currently, the company is also striving to build a zero defect production system according to a quality management system that meets international standards.
Although the company’s quality control process is implemented quite strictly, the implementation in some departments is still not effective. In addition, due to many different factors, the quality of some of the company’s products has not yet reached the set target, the failure rates (FRs) in the process are still high, and there are still unconformable products being moved to the next process and even to customers.
Mold Coil Sub Assy (C23 H5) is a new product line, requiring high precision, very strict technical requirements, and very strict quality standards. Although it has gone through the trial production stage, the product quality is still unstable; there are still many scraps in the assembly and external plastic injection stages.
The current FR of this product is 10.27%. This is a very high rate and is unacceptable compared to the company’s quality system. The company currently has to produce more than the order to compensate for the defect, so the product cost is high and causes a lot of waste. The company wants to reduce the FR to below 6%.
Six Sigma is a popular method widely used in quality improvement. In this article, the application of Six Sigma is discussed to improve the production process of KSVN, identify and eliminate the causes of failure, and minimize waste for the above product line, with the goal of reducing the FR to below 6%.
Literature review & research methodology
Literature review
In the mid-1980s, Motorola pioneered the use of the Six Sigma concept in measuring product quality. In the mid-1990s, General Electric was considered a model company in implementing Six Sigma projects, and the company’s CEO, Jack Welch, used Six Sigma as the company’s popular quality improvement solution. Later, the quality improvement solution based on Six Sigma was applied in many companies, from manufacturing to services such as Texas Instrument, Boeing, IBM, Xerox, Citibank, Raytheon, etc.
Six Sigma evolved from a simple capability measure into an improvement methodology aimed at achieving unprecedented levels of quality by focusing on critical characteristics and identifying and eliminating the causes of defects. Six Sigma aims for a FR of a few parts per million by effectively applying statistical principles and tools for diagnosis, problem solving, and system improvement (1).
Soar and Balanescu (2) introduced the history of Six Sigma as well as highlighted the benefits of applying Six Sigma. In addition, the define measure analyze improve control (DMAIC), design for six sigma (DFSS), define measure analyze design verify (DMADV), identify design optimize validate (IDOV), and difference-in-differences (DIDES) methods were introduced depending on the purpose of use. In which, the DMAIC process—Define, Measure, Analyze, Improve, and Control—was described specifically, and they also introduced how to apply Six Sigma in process management.
Gupta (3) presented a quality improvement study applied at a yarn manufacturing company based on Six Sigma methodologies. He applied the DMAIC process to identify, the cause of the problem and improve quality, reducing the percentage of defective yarn products.
Soković (4) presented a Six Sigma project, undertaken within the company for the production automotive parts, which deals with the identification and reduction of production cost in the deburring process for gravity die-castings and the improvement of the quality level of produced parts.
Ferreira and Lopes (5) used Six Sigma methodology in a semiconductor company to simplify the scrap request process of electronic controllers of wall-hung gas boilers and to reduce its processing time. Through the elimination of problems identified and resolved using the DMAIC procedure, the processing time of scrap requests was substantially reduced, allowing for high cost savings.
Hung and Sung (6) used Six Sigma to solve an underlying problem of reducing process variation and the associated high FR in a food company in Taiwan. The DMAIC phases were utilized to decrease the FR of small custard buns by 70% from the baseline of 0.45% to below 0.141%.
Kurnia et al. (7) determined the level of sigma defects of socks and provided suggestions for quality improvements in reducing the percentage of sock defects. The methodology used in the study is the DMAIC approach. The results of this study are that the level of Sigma increased from 3.7017 to 3.9614, and the percentage of sock defects decreased from 11.08 to 5.54%.
Nguyen Nhu et al. (8) applied Lean Six Sigma to improve garment production processes. The research methodology was based on Lean Six Sigma theory, with the platform of the DMAIC procedure. The tools used in the steps of the DMAIC procedure include cause-and-effect-diagram, Pareto diagram, value stream management, work design, single-minute exchange of dies (SMED), line balancing, Kanban systems, first in first out (FIFO), autonomous maintenance, visual management, design of experiments (DOEs), and control charts. After applying Lean Six Sigma tools, the company had reduced the production lead time by 89.21% from 279 to 30.1 min, reduced the production cycle time by 36% from 25 to 16 s, and reduced the process defect rate by 37.45% from 14.9 to 9.32%.
Research methodology
Six Sigma is a rigorous, data-driven improvement approach with specific problem-solving steps. Six Sigma uses sophisticated statistical tools with a tightly structured problem-solving approach based on the scientific research methodology, including five steps:
1. D – Define
2. M – Measure
3. A – Analyze
4. I – Improve
5. C – Control
The define step identifies the problem, the objective, the product line to be improved, and the scope of the manufacturing process. Tools used in the define step include Pareto charts, suppliers inputs process outputs customers (SIPOC), and project charter.
The measure step collects data to measure the actual capacity of the process through the capacity of the stations. The tool in the measure step is a check sheet to collect data.
The analysis step finds the root causes of the problem, from which solutions are proposed to solve the problem, including the following sub-steps:
1. Identify critical stations with high FRs.
2. Identify root causes in the critical stations.
3. Identify solutions for the root causes.
Tools used in the analysis step include cause effect diagram (CED), Pareto charts, and 5 Whys.
The improvement step implements the solutions identified in the analysis step to improve the process to achieve the goals set in the define step. DOEs is used in this step to optimize the process parameters.
The control step controls the process to maintain improved results. Control charts are tools used to control critical quality characteristics of the process.
The DMAIC procedure would be used to improve the production process of KSVN as presented in the following sections.
Define
Low quality reduced external customer satisfaction and caused great waste for the company. The company had nine product families, including C23, B71, B60, 7PC, V13, CB7, G50, and A01. For product line selection, the Pareto chart of revenue R and cumulative revenue percentage CRP of product families was shown in Figure 1.
Figure 1. Pareto chart of revenues.
The Pareto chart of FR and cumulative failure rate (CFR) product families was shown in Figure 2.
Figure 2. Pareto chart of FRs.
From the two Pareto charts in Figures 1 and 2 above, it could be seen that product family C23 has both the highest revenue and the highest FR. Therefore, product family C23 is selected for further analysis. Product family C23 included many products with CFRs as shown in Table 1 below.
Table 1. CFRs of products in the C23 family.
From Table 1, C23 H5 was the product with the highest FR. This was also a new product with unstable quality compared to the remaining products in the C23 family. This product was required to satisfy strict technical requirements from customers. Therefore, this product was selected for further research. C23H510 0020 included parts and semi-products as described in Figure 3.
Figure 3. Parts in C23 H5.
In order to get an overview and the scope of the manufacturing process of the selected product before starting the steps of measuring, analyzing, and improving the process, the SIPOC diagram is used as shown in Figure 4 below.
Figure 4. The suppliers inputs process outputs customers (SIPOC) diagram of the selected product.
KK was the supplier and also the customer. Inputs include materials such as copper wire, plastic beads, etc., from suppliers, and machinery such as welding machines, wire winding machines, plastic injection machines, testing machines, and packaging machines. The process included 9 stations as in Table 2. The output was C23 H5.
Table 2. Work stations in the process.
From the information defined above, the article briefly describes as in the project charter with voice of customer, critical to quality, objective, and scope in Figure 5 below.
Figure 5. The Six Sigma project charter.
Measure
In order to measure the process capability, the FR of the stations in the process was collected as shown in the following Table 3.
Table 3. The FRs of the stations in the process.
It was found that the process FR was 10.27%, which is quite high. After measuring the current situation, the goal was set to reduce the product FR to below 6%.
Analysis
The FRs were arranged in descending order as shown in the following Table 4.
Table 4. The FRs arranged in descending order.
In order to effectively reduce the FR, stations with high FRs will be analyzed. With the goal of reducing the FR to below 6%, stations of external plastic injection and wire winding would be selected for further analysis.
The external plastic injection station
The types of failure at the external plastic injection station, along with the FR and CFR, were collected and statistically recorded as in Table 5.
Table 5. Failures in the external plastic injection station.
The Pareto chart was constructed from the data in Table 5 as in Figure 6.
Figure 6. Pareto chart for failures in the external plastic injection station.
The CFR of three types of failures, including copper protrusion, mixed materials, and plastic scratches, accounted for 88.4%. These three types of failures were selected for further analysis to find the root cause to eliminate these types of failures.
Copper protrusion
Copper protrusion occurred when the outer plastic thickness covering the coil was not sufficient or the amount of plastic injected was insufficient or uneven. Using the brainstorming method combined with consultation with experts in the production department, the possible causes of this type of failure were presented in a fishbone diagram. Data collection and analysis showed that the two main causes, accounting for nearly 80% of copper protrusion, were as follows:
– Inadequate injection plastic amount
– Improper raw material drying time.
The 5 whys method was used to determine the root causes of the two main causes. Ultimately, the root causes of the problem were as follows:
– High mold temperature during production
– No drying time was recorded on the drying drum.
Mixed materials
Using the brainstorming method combined with consultation with experts in the production department, the possible causes of mixed materials failure were presented in a fishbone diagram.
Data collection and analysis showed that the two main causes, accounting for 62.96% of mixed materials, were as follows:
– Poor mold cleaning
– Inappropriate tools used
Continue to analyze these two main causes to find the root causes. The root causes were as follows:
– Improper gloves
– No mold cleaning equipment.
Plastic scratches
Using the brainstorming method and consulting experts in the production department, the failure of plastic scratches created when taking out the product, the worker collided the product with the mold wall and dropped the product, colliding with hard objects around. The root cause is due to the worker’s incorrect operation.
Wire winding station
The possible causes for failures in the wire winding station included broken wire, scratched wire, improper winding, and some other causes. Among these causes, improper winding was the main cause, causing 90% of the failures in the wire winding station. Further analysis shows that the root causes were due to improper winding tension and the number of turns.
Table 6 summarized all root causes after the analysis step. These root causes would be addressed in the improvement step below.
Table 6. The root causes after analysis.
Improve
After determining the root causes of the problem, solutions to solve the problem were as shown in Table 7.
Table 7. The root causes and solutions to solve the problem.
Improvement solutions were specifically described in Table 7. The DOE solution at the winding station would be implemented to determine the tension and number of winding turns to minimize the FR at the winding station.
The technical parameters of tension T and number of windings N of the wire winding process were given in Table 8.
Table 8. Technical parameters of the wire winding process.
The experiment was performed with two input variables and one output variable. The input variables or factors were tension T and number of windings N. The output variable was the station’s FR. Tension T was tested with seven levels, number of winding N was tested with four levels, as shown in Table 9.
Table 9. Input factors and their levels.
The experiment was carried out with three repetitions in each factor’s combination; the total number of experiments was:
The experimental sequence was conducted randomly. Experimental results were analyzed with the support of Minitab software. The analysis of variance (ANOVA) table was as shown in Figure 7.
Figure 7. The ANOVA table of the experiment.
From the ANOVA table above, with a significance level α of 0.05, we see that the input factors, as well as the interactions between factors, all affected the FR. The number of turns N has the strongest influence, followed by tension T and finally the interaction N*T between the number of turns and tension.
The experimental model was validated with the residual run chart in Figure 8 and the normal probability pin Figure 9.
Figure 8. Residual run chart.
Figure 9. Normal probability plot.
From the graphs, it could be seen that experimental errors varied randomly and followed a normal distribution. Experimental results could be used to determine input variables to optimize output variables.
The contour plot of FR according to tension T and number of windings N was shown in Figure 10.
Figure 10. The contour plot of FR according to T and N.
With the goal of minimizing the FR, the optimal value of the input variables corresponded to the dark blue area in the middle of the chart. This area had the lowest FR (FR < 0.01). This optimal region corresponded to a tension that fluctuated slightly around the value 600, and the number of turns was between the values 628 and 629. The selected values were as follows:
After applying improvement solutions, the FRs before and after improvements at the two critical stations, external plastic injection and wire winding, were estimated as shown in Table 10. The remaining stations had unchanged FRs.
Table 10. The FRs before and after improvements.
It is seen that after improvement, FR decreased from 10.27 to 5.33%, achieving the set goal of reducing FR to less than 6%.
Control
Process control was carried out at the injection and winding stations to maintain improved results.
Control in plastic injection station
At the plastic injection station, the controlled quantities included the plastic injection FR and the plastic layer thickness. The plastic injection FR was controlled by the proportional chart proportional control chart (PCC) 01 with variable control limits. Data were collected, control limits were established, and the PCC 01 chart for controlling the injection FR was shown in Figure 11.
Figure 11. PCC 01 chart for controlling the injection FR.
The plastic layer thickness was controlled by a pair of variable control charts, including sample mean chart mean control chart (XCC) and sample standard deviation chart standard deviation control chart (SCC). XCC was used to control the layer thickness expectation; SCC was used to control the layer thickness variability. Data were collected, control limits were established, and the control charts were as shown in Figure 12.
Figure 12. Control charts for controlling the layer thickness.
Control in winding station
At the winding station, the controlled characteristic was the winding FR. The winding FR was controlled by the proportional chart PCC 02 with variable control limits. Data were collected, control limits were established, and the PCC 02 chart for controlling the winding FR was shown in Figure 13.
Figure 13. PCC 02 chart for controlling the winding FR.
Conclusion
The article discusses the application of DOE to improve the coil manufacturing process of the C23 H5 product in the KSVN company. The research methodology was based on Six Sigma thinking, with the platform of the DMAIC procedure. The tools used in the steps of the DMAIC procedure include SIPOC, project charter, cause and effect diagram, Pareto diagram, DOEs, and control charts. After applying Six Sigma tools, the FR of the process reduced from 10.27 to 5.33%, achieving the initially set goal.
DOE had been applied to determine parameters at the winding station to minimize the FR at this station. Control charts had been used in the control phase to maintain improved results by controlling important process characteristics.
Methods discussed in the article still had many limitations. Improvements were only made at plastic injection and wire winding stations. In addition, the control step used only control charts to control important process variables. Further development includes implementing improvements at the remaining stations to further improve the sigma level of the production process and using statistical tools to control the entire process.
Funding
The research was not funded from any source.
Conflict of interest
The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
References
1. Nhu PN. Six Sigma. Draft2Digital ebook Platform. (2020). Available online at: https://www.draft2digital.com/book/1346390
2. Soar P, Balanescu V. Six Sigma - Strategic Option for Enabling Synergies Within Business Process Management. Business Excellence and Management (2012). Available online at: http://beman.ase.ro/no24/7.pdf
3. Gupta N. An application of DMAIC methodology for increasing the yarn quality in textile industry. IOSR J Mech Civil Eng. (2013) 6: 50–65.
4. Soković M, Pavletić D, Krulčić E. Six Sigma processimprovements in automotiveparts production. J Achieve Mater Manuf Eng.(2006) 19(1). Available online at: http://jamme.acmsse.h2.pl/papers_vol19_1/1836.pdf
5. Ferreira JE, Lopes IS. Improvement of scrap request process with Six Sigma methodology. Proceedings of the World Congress on Engineering 2010 Vol III WCE 2010, June 30 - July 2, 2010. London: (2010).
6. Hung H-C, Sung M-H. Applying Six Sigma to manufacturing processes in the food industry to reduce quality cost. Sci Res Essays. (2011) 6:580–91.
7. Kurnia H, Jaqin C, Purba HH, Setiawan I. Implementation of Six Sigma in the DMAIC approach for quality improvement in the knitting socks industry. Tekstil Mühendis. (2021) 124:269–78. doi: 10.7216/1300759920212812403
8. Nguyen Nhu P, Huu PN, Nam PT, Dang QN. Applying lean Six Sigma to improve garment production processes-a case study. BOHR Int J Oper Managem Res Pract. (2023) 2:53–63. doi: 10.54646/bijomrp.2023.18
© The Author(s). 2025 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.