Model Case & Solution

Today, it is said that printed circuit board production holds more than 90% of assembly process in total electrical device production. For example, soldering procedures for 100 to 2,000 electrical components including 0402 and 0603 chips is as follows.

[Model Case]

- Bare substrates will be fed into printing machine located at left end of line shown below.
- Printing machine will printlace cream solder paste onto bare substrates.
- Inspection machine (SPI) will check solder volume per pad after print.
- A first chip mounter will place smallest components onto substrates at first.
- Next mounter will place biggerlarger chips, IC components, BGAs and connectors.
- Last mounter will place shield cases and large deformed connector parts.
- AOI machines will check mounting states of respective components before reflow.
- Substrate with all components will be put into reflow oven. Solder paste will melt immediately and all the components will be fixed onto substrates at once.
- Another AOI machines will check mounting states of respective components after reflow.
- A finished substrate will come out from machine at right end of line and get flipped to process the opposite side.
* The most speedy line can complete entire procedures of the above in 8 seconds while average line in 20 to 40 seconds.



[AOI Purpose and Requirements]

Mainly there are two way of thinking in terms of AOI purpose and requirements.

-[Quality Assurance]

AOI is to confirm there should be no defect. The purpose is to assure production quality before it is sent to following process. In this case, AOI machines are installed after reflow process and required to detect any possible defect and support necessary repair work.

-[Quality Control]

Another approach is to avoid defect in line by every possible means. By improving production quality to a higher level, defect can be eliminated as a consequence. Usually, AOI machines are installed at steps such as after printing process, before reflow and before IC mounting process. On top of regular OK/NG inspection judgment, they are required to measure mounting status and output information in order to narrow down a possible cause for defect in any.


[Importance of Real-time Inspection]

In today’s speeded-up production lines, incorrect facility setting or wrong feeding of substrate can result in creating a large number of failure parts in a short amount of time. Secondary defect caused by original repair work is another challenge. In order to avoid another occurrence of such defect, it is critical to have inspections for the purpose of “quality assurance” and “Quality control” inline and in real-time. There are 3 critical requirements for AOI machines.
・Detect any error in line immediately and feedback it to upstream process. Do not repeat same error.
・High speed capability to cope with production tact time so that necessary measures can be taken timely.
・Easy and Quick operation so that necessary measures can be taken in real time


[AOI application AFTER reflow process: Pivotal to Quality Assurance]

AOI inspection AFTER reflow is a critical pivot for quality assurance in order to prevent outflow of defects. Therefore, no defect should pass through AOI inspection at this critical step. For example, in case of high-volume, low-mix production, rate of inspection clearance can be possibly improved up to 80-98 % by optimizing production facility and parameters. A number of defect substrate should be limited. In case of small-scale production, production will end before parameters are proved to be appropriate. Therefore, AOI machines are adjusted to judge substrate in doubt as defective parts temporarily and send it to visual inspection by operator to check possible defective area specifically. In that case, a special consideration should be given as AOI machines have a higher defect rate by such operation.


[AOI application BEFORE reflow process: Pivotal to Quality Control]

AOI inspection BEFORE reflow does not have any meaning in terms of quality assurance because it cannot foresee occurrence of possible defects through reflow oven such as floating, solder ball and bridges. Also, each part on substrates will move or pull by surface tension (Self-alignment effect) and possible misalignment or improper soldering defect cannot be caught at inspection before reflow. However, inspection before reflow can check component mounting status and location. Since solder is not melted at this stage, scanning is properly carried out without unstable object on the substrate. AOI inspection before reflow has a very good rate of inspection clearance and low excessive detection. AOI machines will make a warning sound once defect is detected so that operator can visually check subject substrate. Except missing component, defect failures can be repaired by tweezers and when defect rate is higher than what is defined as acceptable level, a warning can be displayed. If a visual inspector should end up repairing same locations repeatedly at this process, he or she can request a staff in charge of upstream machine to check settings. This type of direct information feedback is very effective to improve production quality. There is a case where the line with 40ppm defect rate with AOI machine installed after reflow has been improved up to a few ppm level after moving AOI machines before reflow. Saki has many customers who have the similar experience.


[Inspection After printing]

At inspection after printing process, machines will check skip of solder, stain, bridge, missing, shape defect, volume, mask misalignment and foreign material. Inspection after printing has been performed already for about 20 years. However, it is getting more and more important today due to introduction of 0402 level microchips and parts having bottom electrodes such as QFNs and BGAs. With improvement of mounter precision, over 50% of SMT quality defect root cause is related to volume of solder paste, either excessive or insufficient amount. As solder volume is subject to inspection, 3D measurement is necessary. Recently, as printing stencils are usually cleaned every time prior to printing in the latest SMT lines, printing failure rate is not as high as it used to be. However, 3D inspection machines after printing are necessary equipment in order to review stability by trial printing prior to substrate change, detect any mask cleaning failure and grasp printing condition change. When AOI machines find any failures such as mask misalignment and the generation of strip caused by printing machine’s parameter setting, they can immediately feedback necessary improvement instruction to printing machine automatically.


[Data Preparation]

Inspection data should be prepared in order to perform automatic inspection prior to production start-up. When inspection data is generated, part shape information and pad layout information will be combined to define inspection locations. Usually, part shape information is pulled out from CAD data and mount data with necessary data conversion. Pad layout information can be obtained from substrate design data 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 by combining existing inspection method, using lighting brightness, color and height information. Wizard is available for users as guidance. Inspection data can be registered into library database so that it can be used for different substrate design. Once production is started inline, inspection program should be fine-tuned and optimized at either AOI machines or off-line supporting machines by reviewing mounted component shape or solder volumes. In case multiple lines are in operation, global library function can be used to share updated library data between machines. Machines have a tracking setting function to review determination value based on real production conditions in automatic operation mode.
In case of 2D-AOI machines, since inspection refers to colors of substrate material and component, fine-tuning is required on data after production start or component change. On the other hand, 3D-AOI machines require a very simple data tuning since they use height data for a majority of parameters related to component and solder. It can be handled automatically too. In case of the 3D-X ray inspection machines, the machines pick up only solder joints area subject to inspection after separating top and bottom sides and the data tuning work is very easy.


[Repair Work]

Repair work is to fix defective parts found by AOI machines. There are following 3 patterns.


Use Repair station (terminal) to support operators to locate defective parts on the substrate for repair work.
1. AOI machines read either barcode or 2D code on the substrate and send defect information and images together with ID data.
2. A visual inspection operator will read barcode information on the substrate by a hand-held reader and pull out relevant data at repair terminal to review defect area.
3. OK substrates will be sent to following process, NG substrates to repair station.
4. OK/NG results can be checked by summary menu. It can record who has performed visual inspection and when.
5. By transferring data to WebTracerII, data can be stored for more than one month.



A magnifier micro-scope is installed at repair terminal and an enlarged defect part is displayed on a screen. This will help to save time and labor for locating defective parts on substrates for accurate visual confirmation.



Use AOI machines as magnifier micro-scope for visual defect confirmation.

1. AOI machines will send defect part images to a central management terminal for visual check via network.
2. A special visual inspector will check images and make OK/NG judgment.
(Max. 8 AOIs can be connected to the central management terminal simultaneously.)
3. The inspector’s judgment result will be sent to AOI machine. Only NG judgment result will be sent off to the repair terminal.
4. NG substrates from each AOI machine will be temporarily retreat to NG buffer stocker from line.
In case of Solution2, AOI machines send both real defective substrates and excessive defect determination substrates as NG substrates. However, in Solution3, only real defective substrates retreated to NG buffer stockers will go to repair work process and only OK substrates proceed to following production process. Performed by a qualified inspector, visual OK/NG judgment should be reliable and contribute to reduce assembly man-hours.




[Individual ID management]

In production fields of industrial, aerospace and automotive sectors, each substrate should be given individual ID prior to inspection by reading either barcode or 2D code. Recently, individual ID management has been introduced to productions for smart phone and tablet PC. Each customer has their own operation system. AOI machines are expected to provide appropriate instruction to following assembly machines based on ID as well as output inspection results based on customer’s specific requirements. There are cases AOI machines are required to automatically change inspection data by reading ID or detect any wrong feeding of substrate by inquiring of a server based on ID.


Statistical Process Control (SPC) is a method of quality control which uses statistical methods to improve SMT assembly line quality. SPC server serves to accumulate AOI inspection result data and analyze statistically. It can output production quality report by providing visual graphs for non-defective rates, defective occurring parts, failure types and defective part images based on parameters such as date and production lot per production line. One can tell what kind of defect should be occurring at what location at a glance. When SPC server is hosted online, an operator can access to various data and control them from his or her own PC.


Traceability has been strongly required for production fields of industrial, aerospace and automotive sectors.
Some customers are required to maintain their inspection result data for a long period of time such as a couple of years to 10 years. Once accumulated data in SPC server is saved as long storage data, production history can be checked immediately by entering substrate ID. In addition to defect history and images, entire substrate image can be saved in order to review non-defective components area in past production easily. Also, when repair work history should be made traceable, that would further improve quality of products by assuring traceability.