Applications for inspection

It is said that printed circuit board (PCB) production comprises more than 90% of the assembly process for total electronic device production. Consider the example below of an assembly line where 100 to 2,000 electrical components, including 0402 and 0603 chips, are soldered to a PCB.

  1. Bare substrates are fed into a screen printer located at the left end of line
  2. A screen printer prints solder paste onto bare substrates
  3. A solder paste inspection (SPI) machine checks solder volume per pad after print
  4. The first chip mounter (pick-and-place machine) places the smallest components onto the substrates
  5. The next mounter (pick-and-place machine) places larger chips, IC components, BGAs, and connectors
  6. The last mounter places shield cases and large connectors
  7. An automated optical inspection (AOI) machine checks the condition of the mounted components before reflow
  8. The substrate with the components (populated board) is put into the reflow oven where the solder paste melts, attaching the components to the substrate
  9. A second AOI machine checks the components after reflow
  10. The finished PCB assembly exits the machine at the right end of the line and gets flipped to process the opposite side

The fastest line can complete the entire procedure in 8 seconds, while the average line finishes in 20 to 40 seconds.

Purpose and Requirements

There are mainly two ways of looking at Automated Optical Inspection: purpose and requirements.

Quality Assurance

AOI is used to confirm that the printed circuit board assembly (PCBA) is free of defects. The purpose is to assure that the process and quality are acceptable before the PCBA is sent to the next process. In this case, AOI machines are installed after the reflow process and are required to detect any possible defects and to support any necessary repair or rework.

Quality Control

Another approach is to monitor and avoid defects in the line throughout the assembly process. Taking production quality to this level can eliminate defects. Usually, AOI machines are installed

  • after the printing process
  • before reflow
  • before IC placement

AOI inspects for good/no good components, measures the mounting status and output information, and determines a possible cause for any defects.

Importance of Real-time Inspection

With today's high-speed production lines, an incorrect machine setting or wrong part placed on a substrate creates a large number of failed parts in a short amount of time. Secondary defects caused by the original repair work is another challenge. To avoid another occurrence of a defect, it is critical to inspect for quality assurance and quality control inline and in real time.

There are 3 critical requirements for AOI machines.

  • To be able to immediately detect any error in the line and provide feedback to the upstream process. This is to avoid repeating the same error.
  • A high-speed capability to handle production takt times so that necessary measures can be taken in a timely manner.
  • To be quick and easy to operate so that process adjustments can be made in real time

AOI application AFTER the reflow process: Pivotal to quality assurance

AOI inspection AFTER reflow is pivotal for quality assurance to detect defects that occurred during the reflow process. No defect should pass through AOI inspection at this critical step. Using AOI inspection in a high-volume, low-mix production line to provide feedback to improve and optimize the assembly process can deliver 80-98% improvement.

In small-scale production, production will end before parameters are proved to be appropriate. Therefore, AOI machines are adjusted to look for potential problems and defective parts and send them to visual inspection so an operator can inspect a specific area for possible defects.

AOI application BEFORE the reflow process: Pivotal to quality control

AOI inspection BEFORE reflow is used to check component placement and location. Since the solder is not melted at this stage, the components on the board remain stable. AOI inspection before reflow has a very good success rate for ensuring good quality component placement and problem detection. AOI machines use a warning sound to alert the operator that a defect is detected. The operator can then visually check the board. Except for missing components, detected failures can be repaired by tweezers. When the defect exceeds the defined acceptable level, a warning can be displayed. If repairs are required repeatedly in the same locations, it's a good indication that the settings of the machines upstream in the process need adjusting. This type of direct information feedback is very effective to improve production quality. When Saki's AOI machine was installed and used for inspection after reflow in a line with a 40ppm defect rate, the defect rate decreased to just a few parts per million. Saki has many customers who have achieved similar results.

Inspection After printing

AOI inspection after printing checks for insufficient solder, stains, bridging, missing components, defective shapes, volume, mask misalignment, and foreign material. Inspection after printing has been around for about 20 years; however, it is getting more important today due to 0402 level microchips and parts that have bottom electrodes, such as QFNs and BGAs. Since placement precision has eliminated many placement problems, the root cause of over 50% of SMT defects is related to the volume of solder paste in either excessive or insufficient amounts. To inspect solder volume, 3D measurement is necessary. Solder volume problems have been reduced by cleaning the stencils prior to printing. Saki's SPI machines can be programmed to automatically alert the printer when the stencil needs cleaning.

3D inspection machines after printing are also used to review the printing set-up prior to a board change to detect any mask cleaning failure and identify printing condition changes. When AOI machines find failures, such as mask misalignment or a stripe caused by the printing machine's parameter settings, they can immediately feed back to the printer the instructions needed to correct the problem.

Data Preparation

Inspection data should be prepared prior to production start-up to perform automatic inspection. When inspection data is generated, part shape and pad layout information will be combined to define inspection locations. Usually, part shape information is pulled from CAD and placement data with the necessary data conversion. Pad layout information can be obtained from the substrate design or print stencil production data. As a next step, inspection content and procedures need to be defined for each component. Inspection data can be generated from existing inspection data such as light brightness, color, and height information. A wizard is available to guide users.

Inspection data can be registered into the library database so that it can be used for different board designs. Once production is started inline, the inspection program should be fine-tuned and optimized at either the AOI machine or on off-line supporting machines by reviewing the shape or solder volumes of the mounted component. In case multiple lines are in operation, Saki's global library function can be used to share updated library data between machines. Machines have a tracking setting function to review the determined values based on real production conditions in automatic operation mode.

In the case of 2D-AOI machines, since inspection refers to the colors of the substrate material and component, fine-tuning the data is required after production start or component change. 3D-AOI machines require very simple data tuning since they use height data for a majority of the parameters related to components and solder. This can also be done automatically. Saki's BF-X3 3D automated X-ray inspection machines separate the top and bottom sides first and then identify the solder joints subject to inspection. Data tuning is very easy.


The repair station fixes defective parts found by AOI machines. There are 3 patterns:

Solution 1

  1. Use the repair station (terminal) to support operators in locating defective parts on the substrates that need repair.
  2. AOI machines read either a barcode or 2D code on the substrate and send defect information and images along with the ID data.
  3. An inspection operator reads the barcode information on the substrate using a hand-held reader and extracts relevant data at the repair terminal to review the defect area.
  4. Good substrates will be sent to the next process, no good (NG) substrates to the repair station.
  5. Good/NG results can be checked using the summary menu. It can record who has performed visual inspection and when.
  6. By transferring data to Saki's BF-WebTracerII software, data can be stored for more than one month.

Solution 2

Solution 2 uses a magnifying microscope that is installed at the repair terminal to enlarge the defective part and display it on a screen. This helps to save the time and labor of locating defective parts on substrates and aids in accurate visual confirmation.

Solution 3

Solution 3 uses the remote verification system as a magnifying microscope for visual defect confirmation. This can be accomplished with Saki's software suite.

By utilizing this, it is possible to reduce the visual operator that was placed beside the AOI.

  1. AOI machines send defect part images to a central management terminal for visual check via the network.
  2. A special visual inspector will check the images and make a Good/NG judgment.
    (A maximum of 8 AOI systems can be connected to the central management terminal simultaneously.)
    The inspector's result will be sent to the AOI machine. Only the components judged NG will be sent to the repair terminal.
  3. NG substrates from each AOI machine will be temporarily retired to the NG buffer stack from the line.

In Solution 2, AOI machines send both boards with verified defects and those judged to have excessive defects as NG for repair. In Solution 3, only the boards where actual defects have been identified are sent to the NG buffer stacks for repair work processing and only the boards that have passed proceed to the next production process.

SPC—Statistical Process Control

Statistical Process Control (SPC) is a method of quality control which uses statistical methods to improve SMT assembly-line quality. An SPC server accumulates AOI inspection data results and analyzes them statistically. It outputs a production quality report based on parameters such as date and production lot per production line. It contains graphs of non-defective rates, defective parts that keep occurring, failure types, and defective part images One can tell at a glance what kind of defect is expected to occur at what location. When the SPC server is hosted online, an operator can access the data and control the parameters or assembly operation from his or her own PC.

Individual Identification (ID) management

In industrial, aerospace, and automotive production, each substrate should be given an individual ID prior to inspection by reading either the barcode or 2D code. Recently, individual ID management has been included in the production of smart phones and tablet PCs. Each company has its own operating system. AOI machines are expected to communicate with other equipment in the assembly process based on ID, and then output the inspection results based on customer-specific requirements. In some cases, AOI machines are required to automatically change inspection data by reading the ID or to use ID data to detect whether the components are being properly fed to a particular assembly.


Traceability has been staunchly required in the industrial, aerospace, and automotive sectors. Some companies are required to maintain inspection data results for many years. Once the data that's been accumulated in the SPC server is saved as long-term storage data, production history can be checked immediately by entering the board's ID. In addition to defect history and images, an entire PCBA image can be saved in order to easily review previous production areas that contained good components. When repair work history is traceable, it further improves and assures the quality of products.