Pb-free Design and Process Optimization for PIH
By Harald Fockenberger, Terence Ho, Gerhard Pfennich, Dongkai Shangguan, and Leow Ching Tat
This article examines results of the board design and process optimization for paste-in-hole (PIH) using lead-free solder. The work focuses on the relationship between several design and process variables and PIH solder joint quality. Several important observations were made, including hole-filling behavior, solder joint volume, and reliability.
Strong legislative and marketing efforts are pushing electronics to go green and reduce production costs. Paste-in-hole potentially can meet these needs. While many components are designed for SMT, there are a number of components - particularly connectors - which have through-hole leads due to required mechanical robustness. Wave and hand soldering are the predominant solutions to through-hole component soldering. However, PIH is a potentially more suitable alternative due to cost, process simplicity, and other factors. PIH can reduce production costs by eliminating the wave soldering process.
Several challenges must be solved for PIH technology to be implemented successfully for lead-free soldering. For example, components must be able to withstand thermal exposure of the lead-free reflow process. The assembly process must be optimized by understanding the relationship between PIH solder joint quality and several design and process variables including stencil design (thickness, aperture shape, and dimension); printing parameters (pressure, speed, squeegee angle, stroke, and separation); printing method (squeegee, closed-head); PCB (surface finish, thickness, annular ring design, hole dimension); component (pin penetration, pin dimension, and shape); and solder paste (Sn/Pb and lead-free).
The test vehicle includes a range of components. PCBs were made of FR4 laminate using two surface finishes, i.e. organic solderability preservative (OSP) and electroless Ni/Immersion Au (ENIG or Ni/Au). Three boards measuring 1.0-, 1.6-, and 2.3-mm thick were used for the study. A convection reflow oven with eight heating zones was used with a total heating length of 3.8 m. All reflow cycles were conducted in air atmosphere. Selected tests consisted of solder paste printability, wetting on PCB and component pins, solder balling, hole filling, and pull force.
Stencil design initially was one of the most important issues. The first step was to investigate printing performance of different paste types through the holes. Figure 1 shows the results using a 6-mil stencil. For example, 100% hole filling of a 0.7-mm hole on a 1.6-mm board could be achieved. The increase of paste penetration (length of the paste barrel through the hole) on a 0.8-mm-diameter hole was 0.15 mm. Hole filling after screen printing can be seen from bottom side for different hole sizes (Figure 2).
Figure 1. Hole-filling diagram.
Figure 2. Hole filling (bottom-side view) after screen printing.
Investigating how to apply necessary solder paste volume was the second step. If the calculated paste penetration is larger than 1.5 mm, the aperture should be blocked partly and more paste should be applied on the PCB surface (overprinting). A test was conducted using a stencil with different aperture designs for paste overprinting. A suitable ratio (aperture width to length) should be selected to prevent solder balling. The width-to-length ratio for the stencil aperture for overprinting must be >0.2. Solder and solder-resist type determines maximum overprint.
Different printing methods were evaluated to test hole-filling performance on 2.3- and 1.6-mm-thick boards using a squeegee and closed-head systems from two vendors. A difference was seen using a squeegee for the two different squeegee angles. The 30°- squeegee angle was not sufficient in terms of hole filling, as the volume of solder paste that can be applied in the hole is insufficient and necessary overprint would be excessive. The 45°-squeegee angle performed as expected; however, maximum achievable hole filling was 100% for 0.7-mm holes, below which 100% hole filling could not be achieved.
The closed-head system is expected to have better performance at boards thicker than 1.6 mm. The closed-head system is based on the design approach to control the pressure inside the print head during printing stroke. During printing experiments with the closed-head system, it was observed that on the first few millimeters of the printing stroke, the pressure and paste transfer through the hole were higher than at the end of the printing stroke. This is less pronounced for SMT processes because of low solder paste consumption during one printing stroke. During a PIH printing stroke, however, solder paste consumption is larger. Therefore, it was not feasible to use the closed-head system for PIH printing.
A Design of Experiment (DOE) was conducted on the screen printing process using four different stencils. Stencil thicknesses are 5 and 6 mils, each with calculated aperture size for 100% and 120% paste volumes. The 120% paste volume was used because the transfer ratio for a common screen printing process generally is less than 100%.
Table 1. Summary of solder pastes used.
Two solder pastes were used (Table 1). Paste 1 classifies eutectic Sn/Pb and Paste 2 classifies lead-free Sn/Ag/Cu (SAC). PCB thicknesses were 1.0 and 1.6 mm, with OSP finishes only. Previously determined parameters for the screen printer using a squeegee were used:
Results indicate that the lead-free solder paste had a lower paste transfer (by 8 to 10%) than the Sn/Pb paste due to differing paste viscosities. Board thickness did not influence printed paste volume as much as paste type; and the 1.0-mm boards had smaller solder paste volume deviation compared to the calculated volume of the 1.6-mm board.
The best performance in terms of stable printed solder paste volume occurred with the 5-mil stencil and 100% aperture size, while the worst result occurred with the 5-mil stencil and 120% aperture. This result suggests that the aperture size should be kept as small as possible to prevent a wipe-off effect due to the high printing pressure and flexibility of the squeegee plates.
The second DOE focused on automatic insertion (AI) vs. manual insertion (MI) using two different lead-free solder pastes (Paste 2 and Paste 3) and boards with OSP and ENIG surface finishes measuring 1.0-, 1.6-, and 2.3-mm thick. A comparison was made between components with pin-penetration length above and below 1.5 mm. Visual inspection was performed on all boards; and solder joint defects were categorized.
Comparison of AI vs. MI shows a minor influence on general solder quality. In contrast, the surface finish shows a major influence on solder quality due to better wetting of the ENIG finish.
Comparison of different board thicknesses showed that the 1.0-mm board had the lowest defect rate. On all of the boards, and for all the combinations of board thicknesses and surface finishes used, the pin protrusion length of 2.5 mm on the pin-header connector caused unacceptable solder quality. Preferably pin-protrusion length <1.5 mm should be used. There was no detectable difference between the two lead-free solder pastes.
Board Design Considerations
The pin-to-hole ratio (P/H ratio, i.e. pin cross-section area/hole cross-section area) was an important factor for achieving good results. On the test vehicle, P/H ratios between 0.12 and 0.53 were tested. The analysis of the solder quality at the different P/H ratios has shown that the ratio must be between 0.2 and 0.5. P/H ratios below 0.2 require too much paste and overprint, whereas P/H ratios above 0.5 can cause loss of paste during insertion. The size for the annular rings for PIH was determined to be similar to the common wave soldering process. If overprint of solder paste should become necessary, the copper-pad design should be in accordance with overprint stencil aperture design; and the copper area and overprint should hold the same direction to prevent solder balling (Figure 3).
Figure 3. Copper/overprint direction.
The pull-out force for the pins was tested on the same components and pin positions for all boards tested (Table 2). Results have shown that the PIH process achieved comparable mechanical strength on the solder connection to components that were soldered using a common wave soldering process.
Table 2. Pull test results.
Cross sections showed that the quality of the solder connections (hole filling, fillet, voids) were excellent. None of the tested samples was out of specification (per IPC-A-610-D Class 3).
Various PCB variables, stencil apertures, solder paste printing methods and parameters, component types, and insertion methods have been evaluated for the PIH technology for Sn/Pb and lead-free solders. PCBs measuring 1.0- and 1.6-mm thick with ENIG and OSP surface finishes have been tested successfully using a metal squeegee. Overall, PIH can be a suitable alternative to wave soldering when cost, PCB and stencil design, and component selection are taken into consideration.
This article is excerpted from a paper presented at the 2004 SMTA International, Rosemont, Ill.
HARALD FOCKENBERGER, manager, European Technology Group, Flextronics, may be contacted at +43 4262 2644; e-mail: firstname.lastname@example.org; TERENCE HO, plant manager, Flextronics, e-mail: email@example.com; GERHARD PFENNICH, mechanical designer, Flextronics Design, +43 4262 2644-1353; e-mail: firstname.lastname@example.org; DONGKAI SHANGGUAN, Ph.D., senior director, advanced assembly and environmental technologies, Flextronics (408) 428-1336; e-mail: dongkai.shangguan@ flextronics.com. LEOW CHING TAT, SMT advanced manufacturing manager, Flextronics, 603-51016443; e-mail: Ching-Tat.Leow@ my.flextronics.com.