Industrial Router Crystal 3.2*2.5mm 3225 26M (26.000MHZ) 12PF 10PPM 20PPM 30PPM

Photocoupler

With the emergence and escalation of the Japanese nuclear disaster, the decline of the once-dominant nuclear power sector is an inevitable trend. Will another crucial renewable energy industry—photovoltaics—find itself with renewed growth potential?

In this context, industry experts offer a deeper perspective: While the nuclear power sector may no longer be as vibrant as before, the photovoltaic industry is also facing significant challenges. Since China became the world’s largest producer of photovoltaic products in 2007, improvements have been seen across the board. However, it’s critical to transition from government-subsidized market drivers to market-driven ones. Relying solely on government subsidies poses considerable risks; a change or termination of policies could severely impact the industry. Establishing a genuine photovoltaic market system as soon as possible, allowing the industry to operate under market principles, can maximize the benefits of government support while minimizing the damage caused by policy shifts.

All players in the photovoltaic industry—especially leadership—must recognize this truth: Regardless of industry booms, product prices in every aspect of the industry must continue to decline until the cost of photovoltaic power generation is genuinely cheaper than that of thermal power. Cost reduction remains the true driving force behind the photovoltaic industry’s progress.

But how can costs be reduced? Many immediately think of cutting raw material costs like polysilicon. However, after 2009, polysilicon costs have already absorbed much of the downward pressure. Real competition is now intensifying at the level of company processes and quality management. Similar to the traditional semiconductor electronics industry, the first-time yield of photovoltaic products determines both the robustness of the production process and the magnitude of production costs.

Drawing on the experiences of renowned international photovoltaic firms such as First Solar, Ege Solar, and JA Solar, it becomes evident that controlling the yield of PV products cannot be achieved without refined data analysis and decision-making. This is a technical endeavor, yet it remains a weak point for domestic photovoltaic companies. Advanced manufacturing processes require more than just complex data warehouses or modeling; they emphasize practicality. Recently, a new method of data analysis—interactive visual data analysis—has been embraced by leading companies to identify inferior costs, analyze their causes, develop improvement solutions, and optimize production processes, ultimately boosting first-time yield and overall quality while reducing costs. Interactive visual data analysis encourages the use of more graphical tools for two-way interaction between analysts and data through interactive graphics, moving away from reliance on traditional statistical tools. This lowers the entry barrier for its use and offers better insights into vital data information. Below, we’ll briefly explain how to enhance the first-time yield of solar cells using interactive visual analysis with JMP, a widely-used professional Six Sigma and quality management statistical analysis software in the photovoltaic industry.

The basic process of producing simple solar cells can be roughly divided into eight main steps, as illustrated below:

First, after gathering historical data, we aim to quickly identify the best areas for improvement and determine which processes have the greatest impact on improving the overall throughput rate. From JMP's predictive descriptors, we can glean some insights: examining the linear relationship between the actual yield of the eight steps and the overall throughput rate reveals that the actual diffusion yield and actual etching yield have significantly steeper slopes compared to the actual yields of other steps. Further adjusting the actual diffusion yield and etching yield with the mouse results in the largest change in the actual throughput rate, indicating that diffusion and etching processes are the most influential steps in production.

Naturally, we then want to understand what causes the low etching yields. A Pareto chart easily reveals that the main defects in the diffusion process are rework and debris within the diffusion, while the main defects in the etching process are fragments during loading and machine-generated fragments during etching. Thus, we need to focus on and control the causes of these four defects.

Next, we began using more advanced analytical tools like regression modeling and decision trees to find the root causes. While these tools may seem intimidating to some, JMP makes them accessible through various statistical graphs. For instance, when using JMP’s decision tree function for factor analysis, we simply click the split button on the analysis interface to minimize differences within groups and maximize differences between groups, grouping data to uncover valuable insights:

1. Since the decision tree selects four variables—silicon wafer manufacturer, date, wafer lot, and shift—from many candidate variables, it’s clear these are key factors influencing defect rates.

2. Focusing on the silicon wafer manufacturer, it’s evident that the quality issues at two factories are minor, while the other two factories face more significant quality concerns. This is evident from the decision tree graph, where the green area representing high yield rates corresponds to manufacturers A and D, while the red area with low yields corresponds to manufacturers B and C.

3. Regarding the date factor, November 1 was particularly problematic, as none of the four silicon wafer manufacturers met standards. Additionally, November 2 and 7 also saw major issues, with all production batches of Class B from Factory B failing to meet standards, warranting further on-site investigation.

In practice, solar cell technicians can delve deeper into technical reasons and optimize improvement plans using more interactive and visual data analysis methods. Due to space constraints, further in-depth exploration cannot be covered here. However, it is certain that mastering interactive visual data analysis will be a crucial issue for domestic PV companies seeking to enhance quality management and cost control simultaneously.