Evaluating the Impact of Powder Size and Stencils on Solder Paste Transfer Efficiency

Based on earlier research, this research focused on solder paste powder size, room temperature aging and PCB pad and hole design. The research continued to study powder mesh size, as well as stencil surface treatment and stencil foil tension. The purpose of this research is to identify and arrange the variables that can provide the greatest repeatability, transfer efficiency and print clarity improvements for fine pitch printing. The results show that in terms of its impact on printing quality, nano-coatings, powder types and tension are arranged from high to low, and high tension should be further studied to better understand its most important impact.

Previous work has shown:

One kind. Reducing the mesh size of solder powder can moderately improve printing consistency and transfer efficiency in certain hole designs

b. Under certain conditions, reducing the mesh size of the solder powder may shorten the service life of the solder paste and increase process variables

C. When using high-quality, nano-coated templates under standard mounting tension, pad and aperture design are the most important variables to improve solder paste transfer efficiency

This study added the influence of nano-coating and mounting tension on printing quality to the database, and provided new impetus for the experiment by using similar stencils to print unaged Type 4 (T4) and Type 5 (T5) solder pastes. However, one stencil was coated with a commercial polymer nano-coating, while the other was unmolded and installed with high tension. For reference, the "standard" installation tension is about 35N/cm, while the "high" installation tension is 50N/cm or higher.

Theoretically, higher installation tensions will produce more accurate deposits because the foils will not deflect or reverberate during the separation phase of the printing process compared to lower tension foils.

The test tool (TV) selected for printing research is the Jabil Solder Paste Evaluation Board 2, which can be purchased from leading suppliers of virtual components and test kits. The TV shown in Figure 1 provided a very thorough and detailed analysis in our early research, and it was used again in this research to provide continuity of data collection and analysis.

Features include:

•Print failure (PTF) pattern, land size ranging from 3 to 15mils, shapes include round, square and rectangular land, and are defined by copper (NSMD) and solder mask (SMD)

• 0.4 and 0.5 mm BGA patterns.

• The markings on the PCB are etched with copper instead of screen printing with ink to eliminate the stalemate effect of PCB nomenclature

The smallest feature size printed in this study is 6 mils (150 μm), and the area ratio produced by using a 4 mil (100 μm) template is 0.38. The smallest feature size reported is 8 million (200 microns) because the change increases sharply below this threshold, partly because of the lower AR and partly because of measurement errors. The TV used in the test is the same TV used in the previous study. Use the new panel to collect data for each print. The PCB was not cleaned and reprinted. The dimensions, area ratio and theoretical volume of the pads and holes are shown in Table 1.

The template uses the latest technology and is usually used for fine feature applications that require finer powder solder paste. They are cut from pre-installed brand-name stainless steel on a modern diode laser by a high-quality American mold supplier. A template foil is pre-installed under high tension; the other is standard. Then, the template supplier applies a proprietary polymer nano-coating to the standard tension template. Compare the SPI results of each template with the results in the existing database previously studied. Use the test pad to adjust the size of all holes one by one (1:1), and no holes will be reduced.

The test equipment includes the DEK Horizon screen printing machine in the AIM Application Laboratory in Juarez, Mexico, the Parmi Sigma X SPI machine and the ASH video microscope. The test area is climate controlled and can be operated to simulate the global production environment. For these tests, the test conditions were optimized under the conditions of 25.4°C (77.4°F) and 59% RH and recorded twice a day.

The factory is equipped with SMTA-certified full-time process engineers who have more than 50 years of comprehensive experience in the SMT assembly process. Figure 2 shows the DEK Horizon printer and Parmi Sigma X SPI machine prepared by the laboratory manager for the test run. The printing parameters are as follows:

• Squeegee: 14 inches (355 mm) 60° angle DEK OEM

• Squeegee speed: 40 mm/sec (~1.6 inches/sec)

• Squeegee pressure: 10 kg (~1.5 lbs/inch on a 14-inch blade)

• Separation speed: 1 mm/sec (~0.040 inches/sec)

• Separation distance: 3 mm (~120 mils)

• Under the wiping sequence: use DEK EcoRoll wipes and AIM DJAW-10 solvent for wet vacuum drying (WVD). The mold will be wiped automatically before the first printing of each set of five.

Use a special flat tool support block to provide a firm support for the PCB, and use a new squeegee strip for testing.

Modified SPI inspection parameters to improve measurement accuracy. Generally, a measurement threshold of 30-40μm is used in a production environment to eliminate noise generated by PCB topographic features (such as silk-screen marks, trace masks, etc.). Since the design of this TV limits the noise of topographic features, a measurement threshold of 15μm is enabled to improve measurement fidelity and help detect subtle changes in printing behavior.

The tested solder pastes include Type 4 and Type 5 in modern no-clean flux media. Mix, transport and store the paste under the recommended conditions. The metal percentage of T4 and T5 is the same as the previous test (88.5%). However, different batches of solder powder and flux were used in this test.

The input variables in the experiment include:

• Paste type (4, 5)

•Template nano coating (yes/no)

• Foil installation tension (standard/high)

• Pause time between print tests (0, 30, 60 and 90 minutes)

•PCB pad size (6-15mil)

•PCB pad shape (round, square, modified square with rounded corners)

•PCB pad definition (NSMD, SMD)

The output variables include:

• Deposit amount

•Deposit height

• Transfer efficiency (% volume). TE is based on theoretical pore size rather than measured pore size. All templates are cut by the same operator on the same cutting machine with the same material, so it can be assumed that all system errors are equally applied to all templates. The actual aperture size is not measured

Statistics calculated based on output readings include:

•Average (or average)

• standard deviation

• Coefficient of variation (CV or standard deviation divided by the mean and expressed as a percentage). When comparing different SPI data sets, CV is preferable to Cpk, because CV is normalized to the mean and is not affected by the control limit

•According to recognized industry practice, the acceptable standard is that TE is at least 80% of the theoretical pore size and CV is less than 10%

The steel plate printing under the wipe is performed before each set of five printings, but not between the two printings. Obtain SPI readings immediately after each print. The complete operation includes 0, 30, 60 and 90 minute pauses, and it takes about four hours from start to finish.

The test nesting is as follows:

1. Install the mold and squeegee

2. Stir the solder paste and apply it on the template

3. Print 5 boards

4. Start the timer for 30 minutes

5. Remove the template and scraper, leave the paste on the template

1. Print 5 boards

2. Start the timer for 60 minutes

3. Remove the template and scraper, install the template and scraper

4. At the 30-minute mark of the timer, run WVD

5. Print 5 boards

2. Start the timer for 90 minutes

3. Remove the template and scraper, install "A" template and scraper

4. Run WVD erase at the 60-minute mark on the timer

2. Start the timer for 120 minutes

4. At the 90-minute mark on the timer, run WVD erase

6. Remove mold and scraper

Table 2 shows the data management worksheet, which indicates variables, run sequence, board label and SPI file trace information.

There are many aspects to consider when analyzing the collected data. To simplify the analysis:

1. Data has been divided by component type or filling stack

2. Only review the best and worst cases

3. Relative to the main input variables of nano-coating, slurry particle size and foil tension explain the transfer quality of slurry

0.4 and 0.5 mm BGA holes are the first feature to consider, as these devices are becoming more common in the industry and are a common printing challenge for mainstream PBC assemblers. For 0.5 mm BGA, the area ratio of the test car is 0.71, and for 0.4 mm BGA, the area ratio of the test car is 0.62. These are on either side of the recommended 0.66 AR criterion threshold and have historically been considered the lowest AR that should be considered when printing T3 solder paste. T3 was not included in this test because T4 quickly became an industry standard and is usually easy to obtain. T5 was also tested, which is often requested by assemblers, who believe that the smaller powder size is a quick and easy improvement to the small pitch printing process. T5 paste does bring some printing advantages, but it also brings some inherent disadvantages. Purchasing T5 solder paste may pose challenges to the supply chain, and due to the higher surface area/volume ratio, it will increase the variability of solder paste printing and reflow performance over time.

Figures 3 and 4 show the 0.5 mm and 0.4 mm BGA transfer efficiency (TE) or the percentage of theoretical pore size deposited and the associated coefficient of variation (CV). Each data point on the 0.5mm chart represents the average of 3780 deposition readings, which is 84 I/O per device, 3 devices per board, 3 boards per panel, and 5 panels per test. Each data point on the 0.4mm BGA chart represents 16,200 measurements, because the same number of devices each have 360 ​​I/O. Such a large amount of data makes the results highly reliable.

Please note that TE is slightly higher than 100%. This situation is not uncommon and may be due to many factors related to gasket cracking and/or solder paste extraction. These factors are due to the 1:1 hole:spacer ratio and the combination of round gasket and round hole (Figure 5) 3]. The gasket on this device is NSMD (marked as "copper" in our research). Under this definition method, the difference in shape will naturally cause gasket problems, and usually the ratio of 1:1 aperture: gasket is very high. It is susceptible to positioning errors in templates or PCBs, alignment errors in printers, slightly undersized pads or too large apertures.

Figures 3 and 4 show that for T4 and T5 pastes and coated and uncoated templates, all printing results are within our specified 80% TE and CV ≤ 10% limits. Comparing the results of T4 and T5, T5 has almost no advantage over T4. The only exception is the uncoated template on 0.4 mm BGA, which has an AR of 0.62.

More noticeable than the powder size comparison is the effect of the coating template on printing consistency. In each case, the coated template will deposit a similar or greater amount of solder paste with only half the variation. Halving CV is a major improvement in process control.

For printing experts, this data should help make a decision between fine powder and nano-coating to improve mold release and the overall solder paste printing process.​​​​ Nano-coating has a significantly more positive effect on increasing TE and most importantly reducing CV.

With the transition from low-mix manufacturing to high-mix manufacturing, the need to optimize the production capacity of multiple batches of different products is essential to maximize manufacturing output. The overall production line utilization rate...