Saturday, May 26, 2018
Publication Date: 06/1/2007
Archive >  June 2007 Issue >  Special Feature:
Test & Measurement

AOI: Getting the Product Right the First Time
Optima 7300 in-line post-reflow AOI system.

The evaluation of AOI inspection technology falls into two categories: image acquisition and image analysis. Image acquisition comparisons can be simple. Does the AOI system have the ability to acquire images of all of the areas of the PC board assembly where defects might occur? In post placement or pre-reflow inspection, a single top down high-resolution color camera is best.

Landrex's Optima II Caliper uses a single megapixel CCD chip to capture images. A single chip system provides the best spatial resolution since position information is provided from a single matrix of pixels. Most AOI uses three CCD color cameras that must average the position information from a stack of three chips. When measuring the position of fine pitch and small devices such as 0201 discrete chips, with a single chip camera, sub pixel measurement of 1/20th of a pixel can be achieved. Full spectrum LED lighting is also preferable, since the system must distinguish parts and pins from pads and background. This can be difficult when mixed LED or monochrome lighting is used, especially on small dark parts against a dark background.

In addition, when measuring component positions it is important that there is no lens distortion within the camera's field of view (FOV). This is achieved by using a special telecentric lens that collimates the light reflected from the PC board. Without this lens, part positions on the edge of the FOV are distorted and cannot be measured accurately.

In a post reflow application, a combination of top down and angled cameras allow the system to visualize both the component and solder joints. Many single camera systems are available on the market in both in-line and bench top versions. However, these systems cannot see solder joints on many types of components such a J-leaded devices and LCCs (leadless chip carriers). Nor can they see solder balls and solder bridges that occur close to the body of parts. Angled cameras and programmable lighting allow users the flexibility to view solder joints regardless of their position and from the optimum angle to detect defects such a lifted leads.

For post reflow solder joint inspection, full spectrum lighting and color cameras do not add value. Good solder joint inspection can be achieved using simple algorithms to determine the pattern of reflected light from joints and direct imaging with angled cameras is preferable to implied 3-D from single top down images.

Image Analysis
Does the AOI system use inspection algorithms that can find all the defect types that will occur on the customer's PC board? Many AOI solutions employ single algorithms that are prone to false failure and escapes.

In addition, does the system tolerate normal variation? Such as changes in silkscreen, board color and metallization and partially pasted pads. Subtle and acceptable changes in appearance can require entire programs to be re-taught or adjusting limits on many brands of AOI.

Some acceptable variations in imaging that must be recognized by the system software.

Ease of programming. Can an inspection plan be created in a sufficiently short period? Is the programming method easy to use? Will the user get the same inspection coverage and results, regardless of who creates the program? This important aspect of ease-of-use can be achieved differently on different systems. Where lighting and camera angle are critical, such as in the Optima II Extreme post reflow system, a wizard guides users through the process. In other applications such as high precision measurement pre-reflow, the Optima II Caliper decides automatically and in real- time, which algorithms to use; it requires only CAD data and a complete parts library to create inspection plans.

These are important aspects to consider when narrowing the search among the many available suppliers. However, these criteria alone are not sufficient.

SPC: The Key to Success
AOI is not simply a better method of inspection or an alternative to manual inspection. Yes, AOI promises to provide more accurate and objective SMT defect data than manual visual inspection; and in-line AOI also eliminates board-handling, a potential source of defects. Nevertheless, the real value is that AOI collects defect data objectively, quantitatively and continuously. And this means the customer can respond to process drift and repair defects prior to test.

If the customer is to use the data collected at AOI, the AOI vendor must also provide a method to quickly review defect images, to detect serious repeated placement defects and yield trends and a simple means to filter SPC data to determine root causes.

A defect review tool should show the actual defect in a way that makes it simple to locate. It should also provide a fast verification method, whereby false defects can be reclassified. Finally, the defect review tool should be tied to the SPC software. By this means, false defects can be viewed and filtered as a defect category and false defects can be eliminated from yield and defect data.

SPC Analysis methods should also allow for real-time Alarm Conditions, such as "If ‘X'defect occurs ‘N' times within ‘M' boards, then send ‘an Alarm Condition'". Or, "If Yield drops below ‘Y Percent', then set ‘Alarm Condition'". These alarm conditions can be as simple as stopping the machine or generating a "pop-up" message on the AOI screen or may include more advanced methods such as remote terminal messages or text messaging key personnel. Real-time data analysis capability ensures that quality is maintained throughout the build without requiring constant human monitoring.

Finally, the AOI product must include a simple, easy-to-use SPC data analysis tool that allows process engineers to evaluate defect trends and summarize process deficiencies. Many AOI systems collect defect data in a proprietary format or may even generate third party output files such as Access Database. However, few have had the expertise and foresight to develop simple useful data filtering methods and easy-to-interpret graphical SPC data summaries. AOI data often illuminate weaknesses in the manufacturing process that have their root cause outside of the manufacturing floor. A defect such as "Wrong Part" may lead the quality team to improve on incoming inspection, reduce job changes through better planning and improvements in setup verification. Reducing these errors then allow the AOI data to point the user back to the printer or pick-and-place robot or even the reflow profile. Whatever the root cause, the right AOI system will provide the capability to focus on data-based process improvements.

If the AOI system selected is capable of producing accurate and repeatable results, if you can easily review the defects and extract relevant and timely process data from a system, then AOI can quickly become the best source of product quality data and efficient defect repair. It is also an integral data source for continuous process improvement. The most important second level criterion to consider in an AOI competitive evaluation is the system's integrated Defect Detection and SPC analysis software.

For more information, contact: Landrex Technologies, Inc., 309 Laurelwood Rd., Suite 22, Santa Clara, CA 95054 408-567-9801 fax: 408-567-9760 Web:

search login