Monday, June 25, 2018
Publication Date: 02/1/2012
Archive >  February 2012 Issue >  Special Features: Components and Distribution > 

Root Cause Failure Analysis Improves Efficiency
Component present in tape and uncovered by cover tape.

Root cause failure analysis (RCFA) is the process of determining the source of a problem. In many industries, especially those involving complex machinery, it is common to use root cause failure analysis when problems occur. Why is RCFA so important? Efficiency and accuracy. When it comes to the complicated and expensive machines used in the PCB assembly process, downtime equals increased costs, lost profits, missed deliveries, lost opportunities or all of these. Better efficiency is one of the few ways to compete with manufacturers who have lower labor costs.

With so many machines and specifications looking so similar on paper, features that allow end users to improve efficiency in the production process are gaining importance in the selection process. The use of RCFA tools can assist in improving utilization, efficiency and maximizing profitability.

Several systems to improve the production efficiency of the placement machine have been introduced over the years. There are tools designed for the direct improvement of utilization and indirect improvement. Twenty years ago, few pick-and-place machines had quick change feeder trolleys. Why were they invented? To improve efficiency so the operator could change from one product to another faster.
Image of component in tape allows technician to verify component presentation angle.

Intelligent feeders help the operator set up feeder trolleys offline and reduce the need for "buddy system" reviews of the feeder setup. Splicing feeders and tray changers that can have parts replenished without stopping the machine also were created. All of these systems were designed to directly improve machine utilization and are now common in the industry. Other tools such as self-calibrations are indirect efficiency improvement tools. They reduce the chances of an error or the time needed for maintenance, but do not directly impact utilization. One thing these systems all have in common is that they are designed to improve the efficiency of the known or expected downtime.

RCFA deals more with the unplanned downtime due to unknown or unexpected causes. A failure has occurred. The goal is to find the problem as quickly as possible and get the machine back into production. These tools can also be separated into two categories — after the fact and real time. After-the-fact tools include self-diagnostics that can be run after a problem occurs to see if a system is performing properly. Several tools exist to monitor machine performance in real-time such as tombstone detection and nozzle inspection, but these do not provide any insight into the cause of a problem. They only detect that there is a problem. Few systems exist to monitor the machine's performance in real-time and provide useful information as to the cause of a problem. Unplanned downtime could be dramatically reduced with better root cause failure analysis tools.

Finding the Exact Source
The key to improving efficiency when unplanned downtime occurs is root cause failure analysis — determining the exact source of the problem. Knowing exactly what is causing a problem will allow maintenance personnel to focus their efforts on a clear target rather than just following a checklist of possible problems. Many people have had the unpleasant, inefficient experience of talking to phone tech support personnel who run down a checklist of "possible causes" for different problems. But the reason the checklist is needed is that there is no data available about the root cause of the problem.
Images before and after picking. Note that position on nozzle can be seen in the after image to confirm acceptable pick location.

Without knowing the root cause of any problem, maintenance personnel are essentially working while wearing blindfolds. New systems designed specifically to give better insight into problems can significantly reduce the time technicians need to spend diagnosing problems, and improve equipment utilization.

Defect Prevention
One of the most common errors from the placement machine involves missing or incorrectly placed components. There are two main types of defect prevention systems found in typical assembly lines: systems built into a machine to check itself and standalone equipment whose sole function is to check for defects. An example of a system within a machine to check itself would be post-print inspection on a screen printer. This system checks for defects after the printer has completed its task of applying solder paste to the PWB. This "post print inspection" cannot say why there isn't enough solder paste in a certain location, only that there isn't. Wouldn't it be much more useful to know why there isn't enough solder paste on the board? Clogged apertures, dry paste, squeegee pressure and many other problems could be the cause. Similarly, typical placement machines have several features to prevent defects, but most cannot provide insight as to why the defect occurred in the first place.
Image showing solder paste prior to component placement.

Vacuum sensors are commonly used to detect when a component falls off the nozzle or isn't released on the board. Lasers can check for tombstoning, dropped components, and even incorrect or damaged nozzles. But there are times within the production cycle when these sensors cannot provide useful root cause information. If the vacuum sensor reports an error, why did it occur? All the technician knows is that the required vacuum level was not achieved. Was the component not aligned with the pick position? Is there a problem with the vacuum system? Is the nozzle damaged? Is the feeder feeding correctly? Was the component even present at the pick position? All of these possibilities have to be considered, but each takes time to evaluate.

An example of a machine that is used solely to check for defects would be an AOI or SPI machine. These systems all check for problems that were caused by another machine in the line. They detect missing, incorrect or misplaced components, unsatisfactory solder joints, and other problems. AOI machines provide very useful evaluation of the final PCB, but also have limitations.

While the AOI can detect problems, it cannot tell why a component is missing or incorrectly placed. If an image shows a component placed at the wrong angle, why did this happen? Was the feeder loaded incorrectly? Was the production program written incorrectly? Was the component properly oriented in the tape, tube or tray?

The images from AOI are only of the results, not the process. In addition, there typically is only one AOI machine for the entire line so when a component is missing, the operator doesn't even know which placement machine should have placed the component. The key to improving the efficiency of troubleshooting a missing or misplaced part is having actual images to see what happened during the pick-and-place process.

Reasons for Failures
Focusing specifically on the placement machine, there are many reasons why a component may not be placed correctly on the PCB:

  • Damaged nozzle.
  • Feeder malfunction.
  • Component incorrectly packaged (wrong orientation in feeder).
  • Incorrect pick position teaching.
  • Incorrect component data.
  • Improper board support.
  • Components sticking to the nozzle.

Due to the high speed of today's placement systems, it is extremely difficult for an operator to know which of these factors is actually causing the problems. Therefore they are required to follow a pre-defined list of troubleshooting steps to eliminate one possible cause after another. Throughout the process, they are completely unaware of the root cause. In the end, if they haven't found the problem, they just guess or use experience. This process is clearly imprecise and inefficient. Throughout the entire troubleshooting period, the machine is down and the line is stopped.
Nozzle placing component on PCB.

The time required to precisely locate the source of the problem could be dramatically reduced through better RCFA tools. One example is the use of simple pictures taken during the pick-and-placement process. Images of the various stages of the pick-and-place process instantly show the technician what happened and where the problem came from. This is made possible using tiny embedded high speed micro-cameras along with sophisticated image analysis. Along with the visual analysis by the technician, the software can provide automatic alerts when an error occurs and perform some RCFA on its own.
Component placed on PCB correctly.

By capturing a series of images during the pick-and-place operations, the technician can quickly see what is happening. While some of the questions below are the results and not the cause, they still help reduce the time the technician needs to diagnose the root cause of the problem.

  • Was the nozzle damaged when it picked the part?
  • Was the component present in the tape?
  • Was the component in the correct/expected orientation prior to pick?
  • Was the cover tape properly peeled back?
  • Did the nozzle lift the part out of the tape successfully?
  • Does the position of the part on the nozzle indicate that the pick position is good?
  • Was the part present and properly oriented on the nozzle just prior to placement?
  • Was sufficient solder paste present on the pads?
  • Did the nozzle come down enough/too much when placing the part?
  • Did the component release properly when the nozzle moved up after placement?

In the time it takes to view a few pictures, the technician can accurately determine all of these points. A picture may well be worth a thousand words, but here it may save thousands of dollars. Checking just a single item without the pictures would take as long as checking for all of these points with pictures.
Example of micro-cameras embedded into placement head.

While RCFA is not necessarily a new concept , recent technological developments have made it much more affordable and effective. The use of digital cameras has exploded in recent years. Cell phones drove the widespread miniaturization of cameras to the point where it has become practical to embed them into very small spaces, including the placement head of a pick-and-place machine. Cameras are now widely used in situations where it isn't practical or possible for a human to watch something 100 percent of the time. The city of London has 10,000 security cameras located throughout the city. Why? To help law enforcement find information that is simply not available using people on the street because it isn't practical or economical.
Analysis software showing results of all placements on a PCB along with statistical data.

The days when a placement machine moved so slowly that a human could observe the process are long past. The movement of the typical pick-and-place machine is faster than any other machine in the production line, making it difficult to monitor its activity. The placement heads move over a large physical space, making it impossible for a single camera to accurately see what actions are happening. Small, high speed cameras are needed and they have to be able to move with the placement head as parts are picked in one location and placed in another. Watching by eye would be extremely difficult due to the speed and design of most placement machines. A component may appear to bounce after placing, but did it happen because the part temporarily stuck to the nozzle? Or was there too much placement force on the component? The still images can show what would otherwise be impossible to see.

Critical Technologies
The size, field of view, and resolution are all critical technologies for the camera. It must be small enough to fit inside the placement head without adding significant weight which would impact the throughput or design of the placement head. The field of view must be wide enough to capture the component in the feeder and the component in the up position after being picked. At the same time, the resolution must be high enough to allow the image processing software to accurately detect even subtle differences required for accurate analysis.
Table 1. Comparison of root cause fault detection with cameras and doing it manually.

While the images are the key to locating the root cause, the software is also critical in warning the operator of an error and providing analysis that may not be possible by eye. Powerful software analyzes each image to detect when the component is not picked, is not placed, and even calculates the amount of board deflection during placement. Through communication with the placement machine, it can also show trends such as identifying the feeders, nozzles, or placement sites experiencing the most problems. Although it is common for a placement machine to identify a feeder that is experiencing higher than normal errors, nozzle and placement errors are not as easy for typical placement systems to identify. The analysis of the images is what makes these functions possible or more accurate. This information can also help the operator determine the root cause of problems more quickly.

Finally, the software must be accurately synchronized with the movement of the placement machine. Images must be captured in very high speed at very specific points in time; the exact moment the nozzle goes down to pick the part, the moment the head goes down to place the part, etc. This requires synchronization and very high speed communication between the camera software and placement hardware.

New tools are invented constantly to "make life easier". In a production environment, this often translates to "more efficient". Root cause failure analysis is a tool that can be used in many situations and industries to improve efficiency of troubleshooting. Technicians can quickly find the source of problems so they can be fixed quickly and fixed right the first time. These useful tools are now available for the high speed SMT assembly line with the same benefits: clear and quick understanding of the root cause of the problem, which leads to improved efficiency in fixing the problem. While some sensors and self-diagnostic tools are helpful, they generally do not focus on finding the root cause. Root cause failure analysis tools such as embedded micro cameras to show the exact details of the pick-and-place process along with powerful analysis software can save a huge amount of time versus methods commonly used today and dramatically increase efficiency and equipment utilization.

Contact: Juki Automation Systems, Inc., 507 Airport Blvd., Morrisville, NC 27560 919-460-0111 Web:

Add your comment:

Full Name:

search login