-
- News
- Books
Featured Books
- smt007 Magazine
Latest Issues
Current IssueComing to Terms With AI
In this issue, we examine the profound effect artificial intelligence and machine learning are having on manufacturing and business processes. We follow technology, innovation, and money as automation becomes the new key indicator of growth in our industry.
Box Build
One trend is to add box build and final assembly to your product offering. In this issue, we explore the opportunities and risks of adding system assembly to your service portfolio.
IPC APEX EXPO 2024 Pre-show
This month’s issue devotes its pages to a comprehensive preview of the IPC APEX EXPO 2024 event. Whether your role is technical or business, if you're new-to-the-industry or seasoned veteran, you'll find value throughout this program.
- Articles
- Columns
Search Console
- Links
- Events
||| MENU - smt007 Magazine
Step 2: Process Control
December 31, 1969 |Estimated reading time: 7 minutes
By Phil Kazmierowicz
Technological development, equipment features and capabilities, and material properties have all reached a high level of quality and reliability in the electronics assembly industry. With the industry's focus now shifted to the processing function, a manufacturer's ability to develop and control the individual assembly processes makes a significant difference in producing the required quality at the lowest cost. Zero-defect production is possible where real-time process data provide immediate feedback on the "health" of the process, allowing operators to address a developing problem before an out-of-spec situation develops.
The current industry focus is on how factories can run more efficiently. It would appear that regardless of the equipment on hand, companies have two goals in mind: to reach the required quality and to achieve it at the lowest possible cost.
Today, process control speaks to both objectives. In the past, however, it was more of a tradeoff, i.e., improve quality at extra cost or lower costs and sacrifice higher quality. With modern process control tools, one can experience lower scrap rates, lower rework costs, more uptime, increased productivity and fewer warranty costs.
Process Control Defined
Historically a key aspect of process control recalls the time when companies sought to increase quality by focusing on defect detection to ship only defect-free products. Now, the "ultimate" in control is to monitor various processes continuously and to look for deviations from the ideal. When potential problems are emerging, the information is received in real time so that the process can be adjusted immediately to the optimal situation.
Process control is the ability to acquire data on a particular operation that influences end results. The concept seeks to collect data on a given process and to measure that information against the specifications of the job, and to take immediate corrective action prior to an out-of-spec situation arising.
Process Control vs. Machine Control
Today, assembly equipment is sophisticated and, in some cases, self-monitoring. While very useful, there is a difference between monitoring a machine and monitoring the process output. A case in point may be the reflow oven, which has numerous thermocouples embedded in the heaters and an encoder to control the conveyer's speed. For example, if the heater is set at 200°C and the temperature starts dropping below this level, the thermocouple will detect the difference and "tell" the oven controller to increase the heat output. However, this is not the actual process information.
The temperature profile of a printed circuit board (PCB) going through the oven may vary based on the board's mass, the conveyor's speed, the different temperatures in the oven zones, the airflow, etc. Hence, it is important that actual process data is monitored and not only the machine control data; otherwise, this is simply machine control and not true process control. In the reflow process, this means monitoring the thermal profile of every production board.
Process Development
It is surprising to see how many companies do not properly define their process. Process control without proper process development is useless, so let us take a look at how a process is defined and optimized.
When measuring how a process performs in relation to the specifications, one must first determine the process specification. For the solder reflow thermal process, for example, there typically are two sources:
- The solder paste supplier. There are a wide range of process specifications depending on the alloy and the flux. The required process specification can be found on the paste supplier's Web site and also is built into the software of some of the modern thermal profilers.
- The component supplier also should be consulted because certain parts have even tighter specs than the solder paste.
A typical solder reflow process specification will include a maximum rising ramp rate (3.0°C per second), a flux activation time or soak time (50 to 90 seconds between 140° and 170°C), a time above reflow (40 to 75 seconds above 183°C), and a peak temperature (205° to 225°C).
One must have a well-defined specification before process control can be attempted. However, simply finding an oven recipe (zone set points and conveyor speed) that will give a thermal profile that is within spec may not be good enough. If the resulting thermal profile is very close to either an upper or lower limit, the process will be unstable. As the process is inherently dynamic, even a small process drift may create an out-of-spec situation quickly.
Instead, an oven recipe that will start the process in the center and far from the specification limits must be determined. This optimized equipment setup can accommodate more variables without risking an out-of-spec condition. Today, automatic software for the reflow oven can, within seconds, identify millions of alternative oven recipes and choose the single most optimal configuration.
Defect Discovery Is Not Enough
Statistical process control (SPC) is one technique for process control that was invented in 1924 by W.A. Shewhard. At that time, the tool used techniques to take a few random samples in hopes of representing the manufacturing process. Modern process control methods use sensor technology, which can provide continuous and real-time information on every product being processed.
In the past, when people spoke of improving quality, they often did it by inspecting a finished product. If not right, it was either thrown out or sent back for repair and rework. What eventually went out to the customer was (hopefully) defect-free. Though customers may have been receiving fewer products with defects, obviously this was not the answer. Production is not improving when the manufacturer is producing the same amount of defects, even though he may be able to weed them out before shipping. Costs have not improved because high scrap and rework remain.
Today, process control in its best form actually will monitor the process in real-time and measure actual quality performance of the process. By monitoring process variations, the engineer can adjust the process when it starts to drift out of control. In its best form, adjusting the process quickly based on real-time data will act preemptively to avoid the defects in the first place. Some screen printers have a post-print inspection system. An automated optical inspection (AOI) machine can be put anywhere on the line to catch defects, and systems that continuously monitor the thermal profile in solder reflow ovens are becoming popular.
A Need for Urgency
In the past, true process control has been expensive and requires internal resources to implement. Most companies choose to live without it rather than spend their resources to improve their processes. However, the new technology has made the implementation of process control much less painful and easier to achieve. Process control can improve a company's overall health and bottom line. Many suppliers stand ready to provide helpful manufacturing data for whole-factory implementation. Others supply data for more single process needs, e.g., for the reflow soldering process. Whichever is selected, increasing yields and quality via true process control is the way to stronger, more profitable companies.
Phil Kazmierowicz may be contacted at KIC, 15950 Bernardo Center Dr., Suite E, San Diego, CA 92127; E-mail: phil@kicmail.com; Web site: www.kicthermal.com.
The PWI Concept Allows for Improved Automatic Process Control
The process window index (PWI) is a measure of how well a process fits within user-defined process limits (Figure 1). The following example is done by ranking process profiles on the basis of how well a given profile "fits" the critical process statistics. The center of the process window is defined as zero percent PWI, and the extreme edge of the process window as 99 percent PWI. A PWI of 100 percent or more indicates that the profile will not process product in spec. A PWI of 99 percent indicates that the profile will process product within spec, but it is running at the very edge of the process window. A PWI of less than 99 percent indicates that the profile is in spec and tells users what percentage of the process window they are using: For example, a PWI of 70 percent indicates a profile that is using 70 percent of the process spec. The PWI tells users exactly how much of their process window a given profile uses, and thus, how robust that profile is. The lower the PWI, the better the profile. Figure 2 shows how the PWI is determined. The PWI for a complete set of profile statistics is calculated as the worst case (highest number) in the set of statistics. For example: if a profile is run with six thermocouples, and four profile statistics are logged for each thermocouple, then there will be a set of 24 statistics for that profile, and the PWI will be the worst case (highest number expressed as a percentage) in that set of profile statistics. Note that Figure 2 shows the user designated critical statistics for a single thermocouple.
Figure 1. The PWI, shown as a "bullseye" index, illustrates how well a process measures in user-defined process limits.
null
Calculating PWI
To calculate the PWI: I=1 to N (number of thermocouples); j= to M (number of statistics per thermocouple); measured_value [i,j] is the [i,j]th statistic's value; average_limits [i,j] is the average of the [i,j]th statistic's high and low limits; and range [i,j] is the [i,j]th statistic's high limit minus the low limit (Figure 3).
Thus, the PWI calculation includes all thermocouple statistics for all thermocouples. The profile PWI is the worst case profile statistic (maximum or highest percentage of the process window used), and all other values are less.
Figure 2. An illustration of how a PWI is determined.
null
Figure 3. The PWI formula for thermocouples.
null