5 Steps to Successful Lead-Free Soldering: Step 3.

This article is one of five that describe how to convert tin-lead reflow and wave soldering processes to a more environmentally friendly lead-free process. In the first article (April 2001), materials and machine configuration parameters were discussed. In the second article (May 2001), Taguchi experiments and how they contribute to the development of a lead-free process were explained. This article continues with the knowledge gathered by the experiments and focuses on making the process more robust and, consequently, less sensitive to variation.

Developing a Robust Process

The measure of a soldering process' robustness is its ability to maintain a stable output (yield) with variable inputs. Input variations are caused by "noise" factors. Several factors will vary in a surface-mount assembly even before the printed circuit board (PCB) enters the reflow oven.

First, variation exists in the materials used in the process. Included are variations in: solder paste properties such as constituents, printing properties, activators, lubricant, powder and oxides; board materials, considering different suppliers and different storage properties; and components. Second, variations can occur in the first part of the surface-mount process: solder paste printing and slump and component placement. Third, noise factors can result from the ambient conditions-temperature and humidity-in the manufacturing area. These input variations require the optimum heating profile, which must be minimally sensitive to all variations, and a means to quantify the process capability.

Reflow Profiling

Regarding reflow soldering, the use of lead-free alloys directly impacts process temperatures and, consequently, the heating profile. Increasing the melting temperature narrows the process window because the time above liquidus and the maximum allowed temperature, 250 degrees C, (to prevent component damage and board delamination) do not change.

Triangle (ramp to spike) profile

We can differentiate between processes that are critical and close to the limits of practicality in reflow soldering and processes that are less critical. For processes where the PCB is relatively easy to heat and the components and board materials have temperatures close to each other, a triangle (ramp to spike) profile can be applied (Figure 1). Triangle profiles are intended for products, such as a computer motherboard, that have small deviations in temperature of the assembly (small deltaT).

The triangle profile offers some advantages. For example, if the solder paste is properly formulated for a lead-free triangle profile, shinier joints and improved solderability will result. However, flux activation times and temperatures must be compatible with the higher temperatures in a lead-free profile. The ramp rate of a triangle profile is controlled over its entirety and, during the process, it remains more or less the same. The result is less stress in the PCB materials during soldering. The energy costs are also lower as compared to a traditional ramp-soak-spike profile.

Ramp-soak-spike profile

Smaller components ramp to temperature quicker than larger components and heat sinks. Therefore, to meet the requirements for time above liquidus for all components, a ramp-soak-spike profile is preferred for these processes (Figure 2). The purpose of the soak zone is to reduce deltaT.

Several areas in the ramp-soak-spike profile may result in too much stress in the materials if not controlled properly. First, the preheating rate should be limited to 4 degrees C/sec. or less, depending on specifications. The flux component of the solder paste should be formulated for the profile because a too-high soak temperature can damage the paste's performance; enough activators should be left in the peak zone where oxidation is most critical. A second temperature ramp rate, with a typical limit of 3 degrees C/sec., appears at the entrance to the peak zone.

A third part of the profile where particular attention should be paid to minimizing stresses is the cooling zone. For example, the maximum cooling rate for a ceramic chip capacitor is -2 to -4 degrees C/sec. Therefore, a controlled cooling process is required because the reliability of specific materials and the solder joint structure is also affected.

The best profile for any process can be defined by a Taguchi experiment, as explained in Step 2 (May 2001) of this series. Using noise factors in the experiment will help define which profile is more robust and least sensitive to variations.

Evaluating the Process

The Taguchi experiment produces the best settings for the process. In Step 3, we begin with the settings from the experiment and run the first batch of product through the machine. We now want to quantify the stability of the soldering process over a longer period.

Statistical process control (SPC) is used to stabilize and keep the process in control. In soldering, SPC is used to reduce variability and improve process capability. Typically, x-bar-range charts and capability analyses are used for this purpose. An x-bar-range chart is a graphical representation of statistical calculations on a measured variable, wherein the average and the range (maximum-minimum) per subgroup are used to monitor changes in either the average or the range; the range is used as a measure of variation. Statistically significant changes may indicate process shifts, trends, cyclic patterns or out-of-control conditions due to special causes.

When the most influential parameters of the soldering process, as defined by the Taguchi experiment, are subjected to SPC, improvements in the stability and capability of the process can be easily achieved. For example, in a soldering machine, the hardware and software have been designed to keep important parameters between specified limits around the setpoints. However, even when a parameter is within its deviation limits (no alarm is generated), it can already be statistically out of control or showing a pattern that is unexpected due to historical data.

Buying only hardware and software will not necessarily result in successful SPC. A key consideration is reduction of variability, with a distinction between special cause variation and common cause variation. Control charts are used to eliminate special cause variation, which is any variation that can be linked to assignable causes. Capability charts are used to reduce common cause variation, which is any variation inherent to the process and can only be reduced through changes in the process.

In a reflow soldering process, typical parameters for SPC include the conveyor speed, the gas or heater temperature, time above liquidus and the maximum peak temperatures. In a wave soldering machine, typical parameters include conveyor speed, contact time, preheating temperature (PCB or heater) and the amount of flux applied to the PCB.

An example x-bar-range chart shows the hot air temperature of a preheat zone in a wave soldering process (Figure 3). During an entire day, a sample reading was taken every 10 seconds and divided into subgroups, with five samples per group. The averages and the ranges are shown in Figure 3. The average temperature was 120.0 degrees C, with a setpoint of 120 degrees C. The data originated with a thermocouple that was mounted in the preheat module. All machine data from the process, settings and measured values were logged. The management information file can be imported into SPC software, which will generate statistical graphs such as Figure 3's x-bar-range chart and capability analysis.

We accept preheat temperatures (hot air) of 120 degrees C plus/minus 2 degrees C because we know that, as long as the measured temperatures are within these limits, the board temperature will not fluctuate and will remain within the flux specification. The data, together with the upper control limit (UCL = 122 degrees C) and the lower control limit (LCL = 118 degrees C) will return a process capability (Cp) value:

where Cp = process capability

S = standard deviation.

The Cp chart in Figure 4 shows that, with respect to the preheat temperature, the process is capable. We find the Cp equals 3.55; a Cp value greater than 1.66 is required for a stable process.

Robustness and stability

Once we run the first batch of product with lead-free solder, we need to quantify the stability of the process. These response factors can be measured on the product, like counting defects or data collected from the machine settings. For example, the temperature of a board can be measured with thermocouples mounted on the PCB, or the hot air temperature can be measured in the machine, which is related to the temperature on the PCB.

Another way to measure stability is with special calibration tools carrying instrumentation that ride the conveyor through the machine. The advantages of using these tools are that they are very robust, and several different parameters can be measured in one run. In most production lines, operators have their own test board with thermocouples attached to it. Running the board in the oven, or wave soldering machine, will quickly damage the test board because temperatures are high in lead-free soldering. The board will begin delaminating and warping, and the thermocouples may become detached from the surface.

The parameters that affect quality the most are known from our knowledge of the tin-lead process and from the results of the Taguchi experiment (Step 2, May 2001) on lead-free. We begin counting and collecting the data of these parameters. After the SPC has proved that one parameter is under control (Cp > 1.66) for a longer period, the measuring interval can be reduced. With SPC, only focus on a few, most important parameters. Pareto charts will also help define those parameters to measure to keep the process stable.

The x-bar-range charts indicate process shifts, trends, cyclic patterns or out-of-control conditions due to special causes. In some cases, preventive actions can be taken before out-of-control conditions actually occur.

Exhaust and Temperature Conditions

The entire process has changed with the introduction of lead-free alloys. Temperatures have increased in all modules of the machine. For reflow, higher zone and peak temperatures are the result. For the cooling zone, a more efficient cooling method than normal is required because peak temperatures are higher. Ovens are designed to meet these higher temperatures, but, at this stage of implementation, the machine temperatures should be verified.

Lead-free solder paste has a different chemistry than traditional tin-lead formulas. Therefore, we have to deal with other residues evaporating at different, higher temperatures. Thermal gravimetric analysis can help define where and at what temperature materials evaporate. A sufficient flux management system is needed for a controlled removal of all residues. Additionally, the exhaust and exhaust settings should be verified before implementing the process in the production environment.

Evaluating Reliability

Reliability tests should be conducted to predict the life cycle of the products and compare the data with the standards of the tin-lead process. Shear, pull and thermal cycle tests will tell more about the strength of a lead-free soldered joint. Cross sections will indicate more about the thickness of the intermetallic layers and growth, which are related to reliability.

Release for Implementation

Now we have reached the next milestone in the implementation phase. Once all conditions have been satisfied, we can establish the following:

- the process is stable and repeatable

- machine conditions are under control

- solder joint quality and reliability are within the specifications

- costs are still acceptable.

Therefore, the process can be released for implementation. Thus far, the tests have been conducted on a stand-alone machine, or on a machine in a laboratory or demo room, or on the production machine during downtime. The next step is to transfer this technology to the production lines. However, before beginning production, much work remains to be done. The action items include: project time schedule; quality issues; out-of-control action plan (OCAP); and operator training.

Project time schedule

Create a time schedule for all implementation activities. This schedule will incorporate purchasing materials and additional machine parts if necessary, organizing people and materials to make adjustments, writing procedures and the OCAP, and training the operators and engineers.

Quality issues

The solder in the solderpot (wave soldering) will be contaminated after a longer period of production. Try to establish specifications for the maximum allowed contamination of the alloy. Custumer specifications or guidelines from institutes can help define the maximum allowed percentages of elements in the alloy in your process. In some lead-free processes, these limits are exceeded after 20,000 boards, resulting in the exchange of the solder, which is very cost effective.

OCAP

A process that is disturbed by a special cause variation will show up in the x-bar-range charts. Most operators are very skilled and will quickly recognize this instability. Rapid feedback is possible when the operator controls the process stability. A quick response is necessary to affect the product as little as possible. To keep the process stable, the following steps are required:

- regularly measure the parameters

- verify after every measurement that the process is still stable

-if the process is stable, continue without action; if not, determine the cause of the disturbance according to the OCAP.

Operator training

Line operators should be prepared for the new process. Their training should include instructions on new machine options, different parameter settings (experiences from the Taguchi tests), changes in solder joint shape, dullness and other quality issues. Operators should be trained how to use SPC charts and how to deal with the OCAP.

Conclusion

With the analysis and data from the Taguchi experiment (Step 2), we are able to design a stable, lead-free soldering process. The first batch of products is soldered in Step 3 and, once the product quality is acceptable and the process is found stable, the process is released for implementation. This step is the beginning of many activities that must be completed before lead-free soldering is formally implemented, which will be described in the next article.

---

The remaining two steps will be discussed in individual articles in the next two issues of Circuits Assembly.

---

Bibliography

DeKlein, F.J. (1998). Statistical Process Control for Soldering. Oosterhout, The Netherlands: Vitronics Soltec.

Hall, W.J. (2000). Cooling Parameters in Reflow Soldering. Stratham, NH: Vitronics Soltec Corp.

Kelley, B., and Weisgerber, J. (2000). Step-by-Step SMT Process Control. SMT 2000.

Rooks, B. (2000). Robust Reflow Profile Design.

Suraski, D. Reflow Profiling, The Benefits of Implementing a Ramp-to-Spike Profile. AIM.

---

Gerjan Diepstraten is a senior process engineer with Vitronics Soltec BV in The Netherlands; e-mail: [e-mail].

http://www.circuitsassembly.com

Copyright [copyright] 2001 Miller Freeman LLC