Reading time ( words)
Consumers continue to expect increased functionality in ever smaller electronic products. However, this demand stretches the rules of design-for-manufacture (DfM) and leaves circuit assemblers to wrestle with the combination of decreasing feature sizes coexisting with larger components. Clive Ashmore, DEK, explains how print process control can increase productivity when large and small components coexist on a PCB.
For most progressive SMT manufacturers, fabricating products with small feature sizes and increased complexity is rarely an issue. Nor is the production of products that deploy larger devices. The problem is when assemblies demand both. In surface mount, the imminent rollout of 0.3-mm CSP devices will push feature sizes below 200 µm, yet large components like RF shields and connectors are still required. To address the issues that frequently arise in screen printing this level of mixed-technology assembly, optimizing the print process is imperative. Developing an experiment to help define and guide the optimization process is a practical way of determining which variables to adjust and control.
Figure 1. Calculating the area ratio of a stencil aperture.Stencil Aperture Area RatioThe primary indicator of potential print process issues with mixed technology assembly is represented by a figure known as the area ratio of the stencil aperture, which is calculated as the aperture open area divided by the total aperture wall surface area (Figure 1). Figure 2. When small and large components appear on one circuit board, choosing a stencil for paste print becomes complicated.Figure 2 shows the issues associated with area ratio using thick and thin stencils in relation to large and small component types. The two component types (large and small) can be printed reasonably successfully using a thin stencil, but the large component is likely to suffer from a lean reflowed solder joint due to insufficient volume of solder paste deposited. Using a thicker stencil results in good paste volume for the large component but causes the smaller component to print poorly, as the paste exhibits a tendency to adhere to the larger surface area of the aperture walls, resulting in soldering defects, a lean or even dry reflowed joint. For instance, if a stencil is 100 µm thick and the aperture opening is a square format of 250 × 250 µm, the resulting area ratio figure is 0.625 (250 × 250) ÷ (100 × 250 × 4). Reducing the stencil thickness to 75 µm naturally decreases the total aperture wall surface area, resulting in an area ratio figure of 0.833. Area ratio on its own is a meaningless value; it’s all about how area ratio affects the release of paste from the aperture. This transfer efficiency is critical to the print process; this figure can be manipulated by adjusting the geometries of the stencil design. The aperture open area is not a variable that can be manipulated at will in the stencil design. It is entirely defined by the size of the pad on the board, which is there to suit the component type to be imaged and ultimately assembled. Table 1. Common SMT devices along with the typical area ratio exhibited by each.Device type Area ratio0603 2.240402 1.5202010.980.75mm CSP0.650.5mm CSP0.60.4mm CSP0.5 0.3mm CSP0.4Notice in Table 1 the smaller area ratio figures attributed to CSP devices, and how that figure diminishes with finer pitch. From empirical trials, it has been clearly ascertained that transfer efficiency and area ratio retain a linear relationship until the area ratio drops to a figure of about 0.66, which correlates to a transfer efficiency of around 75%. Below this point, the relationship becomes distinctly non-linear, indicating a significant reduction in transfer efficiency to less than 75%, which is deemed unacceptable for a reliable joint at reflow. Therefore, when designing stencil artwork, the area ratio rule-of-thumb is set at 0.66 or above (defined in IPC-7525). This is easily implemented if you’re only concerned with one type of component on a board — you can simply use the rule to determine a stencil thickness that offers acceptable transfer efficiency above the 0.66 mark. This is not the case when trying to accommodate mixed technology componentry. Here, the task is to increase the transfer efficiency for apertures that fall below the 0.66 area ratio point, permitting these smaller board surface features to be successfully imaged using a thicker stencil, which simultaneously fulfils the mixed technology requirement with larger components. Figure 3 illustrates how transfer efficiency tails off as feature sizes and their corresponding area ratio figures reduce. Note how the 75% acceptable transfer level correlates to an area ratio of 0.66. The Experiments The objective for the experiments is to capture a series of transfer efficiency values resulting from different sets of print process parameters (variables). Figure 3. Transfer efficiency as related to feature size.Table 2.Aperture size Area ratio1000.2501250.313 1500.375 1750.438 2000.5002250.563 2500.6252750.6883000.750 3250.8133500.8753750.9384001.0004251.0634501.1254751.1885001.2505251.3135501.375The chosen experiment set up requires an automatic stencil printer to apply solder paste through an industry-standard 100-µm-thick laser-cut stainless-steel stencil. The Table 3.Stencil thickness100 micronsStencil materialStainless steel (grade 330)Stencil fabrication methodYAG laserSolder paste compositeLead-free SACSolder paste sizeType 4Solder paste metal loading88.9%ToolingVacuum blockSqueegee holder length150 mmprinting machine, stencil, squeegee blades, board support tooling, solder paste, and operators were kept constant throughout to reduce unintended variation. To capture relevant data sets, the stencil was designed to image a full range of SMT and area array devices. To fully observe the process capability for each experiment, a decreasing aperture array was deployed; Table 2 lists the aperture open size and each associated area ratio. Within any stencil printing process there are many significant factors that directly influence the process. These variables include print speed (squeegee excursion velocity), print pressure, and squeegee angle (attack angle), and even squeegee blade overhang. The squeegee assemblies used for the experiments were 45° attack with a 6-mm overhang; 60° attack with a 6-mm overhang; and a 60° attack with a 15-mm overhang. The expectation was that the 15-mm overhang blades would present a variable resultant squeegee angle depending upon the print pressure. The 6-mm overhang blades were used to contrast and compare the 15-mm overhang results. To ensure the experiment represented a standard set-up, stock materials were used throughout (Table 3). The solder paste deposits were measured using a 3D measurement system. By adhering to a strict methodology, the only process settings varied during the experimentation were print speed, print pressure, and resultant squeegee angle. Table 4 details the print parameters used throughout the experiment, with no other parameters adjusted. Results A noticeable correlation between process set-up and the distribution of resultant transfer efficiency was observed for both area array devices and SMT components. The results obtained from the five process settings for the 60° 6-mm overhang squeegee showed great similarity for all apertures independent of the process settings, meaning that a high-pressure, low-speed pass gives a result that differs less than 3% from those obtained deploying a low-pressure, high-speed set up. This set of experiments produced transfer efficiency results that showed the 75% point correlating to 225-µm diameter features being the smallest that could be imaged reliably. The 45° 6-mm overhang assembly also exhibited a tight distribution independent of the process set-up. Within this dataset, the transfer efficiency was positively affected with the 75% acceptable cut off value now correlating to 200-µm diameter apertures, signifying a process cable of printing area ratios down to 0.5. Transfer efficiency results obtained from the 60° 15-mm overhang squeegee help illustrate the impact of a flexible blade assembly. With this set up, the effect of process parameters can alter the transfer efficiency by a delta of up to 15%. Under high pressure and low speeds, transfer efficiency was significantly improved. These process conditions revealed apertures of 175-µm diameter to be above the 75% point, indicating a process capable of printing area ratios as low as 0.438. Table 4. 60°, 6mm overhang45°, 6mm overhang60°, 15mm overhangExperimentsPrint pressure in Kg Print speed in mm/secPrint pressure in Kg Print speed in mm/secPrint pressure in Kg Print speed in mm/sec1620620620213201320132039.6509.6509.65046806806805138013801380These experiments show that changing the squeegee and print process variables allows operators to influence transfer efficiency. Today’s SMT fabricators need to produce products that have extremely small features alongside large features, but in a single imaging process. These experiments show that fairly simple solutions are available. For mixed fine- and large-feature technology assembly, the best suited short overhang blade is the 45° squeegee set up, which exhibits better transfer efficiency on area ratios while external process influences have little detrimental effect. The 60° 15-mm overhang blade is the most compatible with mixed technology assembly but the process parameters do have a significant impact on the transfer efficiency. Therefore, this solution requires a reasonable level of printing knowledge to fully optimize the process. Clive Ashmore, global applied process engineering manager, DEK, may be contacted at firstname.lastname@example.org; www.dek.com.