Why the need for 3D Solder Paste Inspection?

Ray Welch, Senior Applications Engineer, Koh Young America, Inc.
Brent Fischthal, Senior Marketing Manager, Koh Young America, Inc.

People may use the phrase “This is where I am feeling the pain” when describing who and where an issue is causing problems with their board assembly process. Yet, reversing the phrase into a question like “Where do you feel pain in your process” may help better reveal areas or specific issues where solder paste inspection (SPI) may offer a solution to eliminate the issue.

Where do you feel pain in your SMT process?

Walking through several questions can help guide a discussion about how a focus on paste printing can address the current and potential issues in the process.

1. Are escapes causing field returns with costly rework and customer dissatisfaction?

2. Has low yield or rework issues caused missed delivery deadlines?

3. Do stencil designs require too many time-consuming and costly revisions?

4. Has a substandard print process resulted in missed yield targets and excessive scrap?

5. What is the solder print process capability index (Cpk)?

6. Is the performance data readily available from the print process equipment?

7. Has the absence of SPI caused missed customer or product opportunities?

Of note, is that OEM customers judge the capability of their (prospective) contract manufacturer(s) by their equipment sets and processes. SPI is becoming a more critical decision factor in selecting a new CM, or placement of new business with an existing CM, for product that requires use of SPI due to PCB design complexity and/or small feature sizes, or simply, the prevention of print defects into the SMT process and beyond.

The Power of SPI

Real life customer examples show the value and power of SPI, attempting to answer the question “Why the need for 3D SPI?” Many new customers to SPI quickly realize that their print process is not as robust as they might have thought and is contributing to higher levels and cost of rework, and potential customer escapes. Here is one truncated real-life anecdote.

The Sins of the Process

Experts say, “SPI reveals the sins of the print process.” That has been clear for many customers who are new to SPI. Once they begin inspecting product with SPI, issues in print performance or printing practices may become immediately obvious; be it stencil design (low area ratios), poor printing practices and/or printer setup, poor board support, less-than-optimal print parameters, poor performing solder paste, PCB design or fab issues (excessive silk screen, thick solder mask), etc.

For example, a mid-tier EMS producing a new product used by a Japanese tractor company. Early product was built without the use and benefit of SPI, so print-related issues were not that clear, particularly for the ones that escaped the print process and onto the customer. This resulted in BGA head-in-pillow (HIP) and early field failures.

Several months later, the manufacturer implemented SPI on the lines producing this product. Print issues were immediately identified in the form of insufficient solder defects with paste volumes well below the target range of volumes.

These issues are the probable cause for the BGA head-in-pillow issues. Figure 1 presents the %Volume (Paste Transfer Efficiency) results for the BGA though use of a Minitab Run Chart.

Figure 1: The 3D SPI revealed a serious insufficient solder print issue that was the contributor to the product failures due to BGA head-in-pillow.

Data Analysis

Minitab Run Charts offer the most graphical presentation of print performance, plotting the individual pad %Volume results (light gray dots) for the selected pads – in this case, the 0.5 mm pitch BGA pads that had experienced HIP – and the Average %Volume (blue dots) for the selected BGA for each board.

The results are plotted for each board inspected with a time-based analysis of the print performance, which offers insight on shifts or trends in the results as the build progressed and when defects occurred.

Typical SPI volume tolerances are 60% – 160%, identified as the Target Volume Range in the graph above. Any results below 60% are an insufficient solder defect and potentially classifies the board as a Fail / NG (Not Good).

In this example, virtually every board should have been failed. The same would have been true for boards built several months earlier with the same print process, but without the benefit of SPI. Without SPI, the cause for the in-house BGA test failures and the field test failures was not clearly identified, or even suspected as print related issues.

Having SPI during the early NPI builds would have found the print issues and driven process improvements, before shipping to the customer, avoiding costly rework and customer dissatisfaction.

Implementing SPI helped:

  • Identify a key contributor to the BGA head-in-pillow issue
  • Define print performance issues with the current print process
  • New printer under evaluation
  • Stencil design (BGA area ratio 0.60 vs. target > 0.66)
  • 2×3 Panel layout, creating PCB warpage during first pass reflow
  • Less-than-robust printing practices and operator behaviors

The use of SPI would have detected these issues during the NPI builds, driving process improvement and preventing escapes to the OEM customer. The primary goal of SPI is to prevent print process escapes (internal and external).

Prevent Print Process Escapes

Implementing SPI helped to identify the issues noted above, which included detailed data in the evaluation of the new printer, and a data-driven comparison between two candidate printers. The print performance data collected via SPI helped to justify failing the evaluation of the printer used in this example, and the selection and deployment of a best-in-class printer.

Using SPI during the NPI builds would have helped find the print issues that resulted in the BGA headin- pillow issues, as well as other print defects, resulting in costly rework, process escapes to the customer, and customer dissatisfaction.

The primary goal of SPI is to prevent print process escapes, to both internal (downstream process) and external customers.

Print Process Characterization

SPI provides a detailed view of the print process. Paste Transfer Efficiency, displayed as %Volume (Measured Volume / Stencil Aperture Volume * 100%) is a key measure of the print performance.

The upper Minitab Run Chart plots the individual pad results for a BGA that was experiencing significant Insufficient Solder defects – defects that were not identified prior to implementing SPI, which clearly indicated that virtually every board had print issues.

The cause for the BGA Insufficient Solder defects was a low stencil aperture area ratio (0.60 versus a goal to be > 0.66 for a standard laser-cut stencil). SPI helped bring focus to the stencil design and validated the improvements after correcting (raising) the area ratio for this and other components.

In Figure 2, the lower run chart reflects this improvement, with the %Volume results averaging in the typical range of 100-120%, and the overall results within the volume tolerances of 60-160%. Use of a 5-mil versus a 4-mil stencil, was the only difference between the two charts.

Figure 2: Before and After comparison of the individual pad results for a BGA that was experiencing significant Insufficient Solder defects

Similar before-after comparisons can be made for all component / aperture types, with the smaller and more critical features / apertures showing the greatest improvement.

Such process characterization and validation are only possible through use of the parametric data collected and presented by SPI, versus simply reacting to print defects or process yields. SPI must be used to fully optimize the print process.

Improper Print Practice

During characterization of the print process, issues related to poor or improper printing practices, particularly on the part of the operators, may be identified.

The upper chart presents an issue that is often found during customer visits – not performing paste kneading prints before resuming production printing after a long pause in printing (> 30 minutes, depending on the paste selection and condition), or before beginning printing after a changeover.

It typically takes about 4-5 boards before a stable process is achieved for the condition described above, often with the need to wash and reprint the first few boards. The results shown in the upper chart provide the data that supports the best practice for use of paste kneading prints.

The lower chart in Figure 3 reflects another common issue – only replacing squeegee blades when they are dinged or damaged, not based on blade edge wear. Such wear can be difficult to physically detect. But it will slowly become clear in the print results, often through higher paste volumes and greater print variation, and, through residual solder on the stencil after a print stroke. Similar issues and SPI results may occur for excessively worn or damaged stencils.

Figure 3: During characterization of the print process, issues related to poor or improper printing practices may be found (Top: Solder Paste Kneading, Bottom: Worn Squeegee Blades)

The use of SPI helps to identify such issues, with the parametric results used to support and validate improvements in the printing practices and demonstrate to operators why they are asked to perform such practices.

Print Process Evaluation

The use of parametric data, versus print defects or process yields, is the only means for effectively characterizing, then optimizing the print process – data that is easily collected and presented using SPI.

Equipment and materials require parametric data for evaluation or process validation. SPI provides a powerful tool for evaluating:

  •  Stencil Printers
  • Stencil Technologies and Aperture Designs
  • Board Support and Clamp Tooling
  • Solder Pastes and Process Materials

Such analysis cannot be effectively performed using just defects or yields. We must be able to “see” the full process, by way of pad-level paste deposit volume, height, area, offset, and more.

The upper Run Chart provides a time-based view of the print performance for a new printer under evaluation, presented through the %Volume (Paste Transfer Efficiency) results for each pad of a particular BGA location.

Although the print results started-off more-or-less okay, they soon deteriorated with greater process variation and increasingly more Insufficient Solder defects. The cause of the problem was associated with overheating of the print area, due to poor ventilation of the heat generated by the understencil cleaner vacuum pump. This led to the solder paste increasing in temperature from ambient to over 90F, resulting in paste slumping.

In figure 4, the lower chart shows the same product, stencil and paste run on a best-in-class printer, with a dramatic improvement in print results. Such data, only available through use of SPI, clearly presents the comparative results between the two printers.

Figure 4: A time-based view of the print performance for a new printer under evaluation (top) and best-in-class printer (bottom) for comparative analysis using the %Volume (Paste Transfer Efficiency) results for each pad of a particular BGA location.

Summary

Implementing 3D Solder Paste Inspection serves to:

  • Prevent costly print process escapes to the customer (internal and external)
  • Detect print process issues during NPI builds, driving process improvement and preventing escapes during process and product stabilization
  • Provide a robust tool for the Characterization of the print process, required to fully optimize the process, increasing product yields and reducing rework/scrap
  • Drive implementation and validation of paste printing best practices
  • Provide a powerful tool for evaluating:
  • Stencil Printers o Stencil Technologies and Aperture Designs 
  • Board Support and Clamp Tooling 
  • Solder Pastes and Process Materials 
  • Enable new customer / product opportunities that require solder paste inspection

The above simply describes the benefits of SPI, tying back to where the customer is asked “Where do you feel pain in your SMT process?” A discussion may help to uncover other benefits that play into the customer’s current process or issues.

And again, of note, is that OEM customers judge the capability of their (prospective) contract manufacturer(s) by their equipment sets and processes. SPI is becoming a more critical decision factor in selecting a new CM, or placement of new business with an existing CM, for product that requires use of SPI due to PCB design complexity and/or small feature sizes, or simply, the prevention of print defects into the SMT process and beyond.

Many manufacturers believe a “best-in-class” process requires using an SPI as more than a “Go / No Go” system. It must fully characterize the print process and make the necessary process improvements. Some suppliers, like Koh Young for instance, offer comprehensive on-site training to help customers use SPI more effectively, as well as offer recommendations and best practices for improving the print process and achieve a “best-in-class” process.