Articles

Paul Groome, Machine Vision Products. July 2025

by Paul Groome, Machine Vision Products Published in US-Tech in February 2018

When Machine Vision Products, Inc., (MVP) was founded 25 years ago, cell phones were in their infancy and portable electronics simply didn’t exist. The majority of manufacturing was through-hole, the i486 was the main computing platform and Windows 3.1 was flashy and new. Dr. George T. Ayoub built the company when no one had seen a smartphone, flatscreen TV, mp3 player, solid state drive, GPS system, hybrid car, digital photograph, and many other technologies ubiquitous today. And, who could foresee the impact that the Internet would have?

In 1993, cell phone usage in the U.S. had risen to 11 million users, whereas by 2017, 81 percent of the population, approximately 265 million people, had cell phones. At the time, the Sparc processor — MVP’s original computer of choice — had 0.8 million transistors. Today, the Xeon processor used in MVP’s systems contains up to 7.2 billion transistors. When the company was founded, there were only 50 Internet (www) servers worldwide. Now, more than 50 percent of the world’s population has access to the web.

The rate of change in technology and manufacturing in the electronics industry has been exponential. Famous for his prediction of the doubling of the number of transistors in ICs approximately every two years, Intel’s former president, Gordon Earle Moore, should be proud.

New Challenges Ahead

MVP has always endeavored to provide  innovations, solutions and capabilities to meet the needs of the rapidly-changing market. These include statistical process control (SPC) for AOI, SPI and 3D AOI, multi-spectral lighting, multi-camera inspection, wire bond inspection, and die metrology. In the early 1990s, a system with 1 to 2 mil (25 to 50 μm) resolution may have been suitable to inspect a 486 motherboard. In 2017, 0.4 to 0.6 mil (10 to 15 μm) resolution was enough for basic componen assembly. But, moving forward, resolutions in the single-digit micron range will be required for key electronic hardware.

The next generation of chip components are challenging SMT inspection with geometries of 5\ mil (125 μm) and solder fillets below 1 mil (25 μm). Even thermal expansion on a board can change the position of these components dramatically, so special registration techniques will be required. MVP, with its 2020, 850, Spectra, Supra, and Ultra platforms, is well-positioned to deliver the solutions for this next level of integration.

Into the 2020s

Key goals for any manufacturer are to increase quality at the most critical stages of manufacturing, increase yields and decrease costs. Although the 10 times rule for the cost of identifying a defect at each stage is less accurate as time goes on, the concept is still true.

Using the automotive industry as an example, finding a defect at the lead frame process may cost less than $1. Finding one during SMT assembly may cost between $10 and $300. Finding a defect in the field can cost between $1,000 and $3,000.

Finding defects as early as possible is critical to providing the highest possible quality at the lowest cost. This is true for all aspects of electronics manufacturing. The 10x rule varies by market and while 10x may be accurate for consumer electronics, high-rel products may be 100x and mil/aero 1,000x.

Looking toward the next decade, MVP has expanded its range of systems to meet the future’s new and challenging inspection requirements. The company offers front-end semiconductor process inspection, including wafer, die and post-dice inspection. Back-end semiconductor processes include lead-frame and wirebond inspection, die placement metrology, surface and substrate inspection, packaging inspection, and hybrid and MCM inspection. For SMT assembly, the company offers backplane, solder paste, post-placement, post-reflow, and conformal coat inspection.

Front-End Processes. The latest series of 850 platforms provide fully-automated handling of wafers in film frames for surface and post dicing inspection. Surface damage, FM and edge damage can be identified. These systems can be configured for cleanrooms up to Class 100. The company offers fully-automated handling for wafers in ring frames, top/bottom side inspection, defect marking, and defect mapping.

Back-End Processes. The company’s offering for back-end processes are split between two platforms: the 850 and its latest platform, the 2020. These processes cover wirebond, die, edge, surface, FM, eutectic, lead frame, BGA, bump, flux, and paste inspection.

Each solution offers high-resolution tele-centric optics, 3D laser inspection, confocal, or 3D projector imaging. This makes these platforms extremely flexible. In addition, the systems provide capable registration tools, allowing dice, wires and substrates to be registered on the fly, with inspection requirements calculated on a part-by-part basis.

SMT Assembly. For SMT assembly, the company provides solutions to meet nearly any need in the industry. Built on the experience of the company in the front-end and back-end semiconductor markets, MVP can now provide the same techniques, resolutions, and 3D tools to SMT manufacturers.

The company faces new designs, techniques and challenges regularly. Working with customers, the company identifies trends in manufacturing and uses them to find the best place for its defect detection systems. This development strategy is enabling the company to offer solutions for today, tomorrow and the decades to come.

When defining an inspection strategy for your manufacturing processes there are many factors that need to be considered. As well there are many products that seem to have solutions to your inspection requirements, how do you make the decision of which system to use? How do you determine the viability of manual inspection, of template/comparative vs. measurement based AOI, what resolution to use, which cameras, what data outputs are required, do you need measurement data, attribute data or both.

There are many questions that can be asked and the goal of this paper is to detail what variables are important vs. those that can confuse the AOI selection process. For this purpose this paper will discuss the top five variables that can ensure a successful AOI deployment in your manufacturing processes.

Figure 1 – Structural DPMO Rates

♦ 1. Ask yourself, why are you purchasing an AOI system? Defect coverage?

The first question you should ask yourself is why you are purchasing AOI? Hopefully the answer is to detect defects. From previously experiences some users even leave this out from a system evaluation.

When you look at defect coverage do you know your process and the defects that can arise during SMT manufacturing; solder defects (tombstones, opens, solder amounts etc), presence/absence, component measurement data, wrong components etc. This knowledge and how you use this data is important to making an inspection decision. There are many industry bodies that can help you to determine yields and possible defects. As a guide, based on internal studies and industry standards, DMPO (Defects Per Million Opportunities) rates are defined in table 1 and 2, and can be used as a base line for defect detection requirements.

When considering defection inspection capabilities there are a number of items to consider;

Figure 2 – Typical Process Defect DPMO Rates

♦ i, Inspection techniques, capabilities of template matching vs. measurement based AOI platforms.

There are two basic techniques used by AOI systems; template based comparison systems and rules based measurements systems. The main difference is measurements. For example a rules based system will measure the width, length and position of a component then from this position measure the actual amount of solder on each pin (referenced to component position), then perform OCV or OVR dependant on component marking. The component body measurements provide the capability to find the correct component and the solder measurements will provide the exact percentage of solder between the component and the pad. Whereas a template based system will look for presence absence of the component and compare the joint to a known good joint, which can lead to a higher volume of false calls. When, as in most manufacturing scenarios, defects will follow the DPMO rates shown in Table 2 and you will have solder defects, tombstones, and opens it is important to select a system that has the capabilities to inspect for solder.

Another question that is normally discussed is do we need to use angled cameras to locate defects. There are many arguments to this, but generally angled cameras are complex to calibrate and can be significantly affected by board warp. If you understand the reflow process, a single telecentric camera system can provide the same defect coverage as an angled camera without calibration or warp issues. Figure 1 shows a real life example where the left center pin (pin 25) is open. On a single, telecentric camera system using tri-color lighting (Tri-Color technologies use angled lighting vs. angled cameras) the defect is easily visible by inspecting the changing reflow profile. As you see in the example the failing joint shows up red from the red 45 degree lighting vs. the high green lighting which shows good joints.

Figure 3 – Jlead defect on a single camera system using Quad Color Lighting

♦ ii, Detection Toolbox, what tools/algorithms are available to detect defects.

As technology changes are a constant in our industry, it is important to understand the flexibility of the inspection tools and future capabilities of the AOI system. Is the system always going to be post reflow? Will you ever look to put the system in a new position in the line? What technologies are you building today and what is your future roadmap for manufacturing?

Some AOI systems are purely focused at a single task i.e. Paste, Pre-Reflow, Post Reflow or Post Wave. More experienced AOI suppliers will provide solutions that cover all aspects of the manufacturing line in the same system, provide integrated 2D/3D technologies, while also providing capabilities for microelectronics and semiconductor inspection. If you are looking at maximizing the ROI (return on investment) for your AOI purchase, process position flexibility and inspection capabilities of the system are key to the success of your investment. As an example an experienced AOI supplier such as can provide extensive capabilities (shown in Figure 2), which allow for many different inspection scenarios beyond basic SMT inspection.

Figure 4 – Inspection Techniques

♦ iii, False call rates and how these affect defect coverage.

For fault coverage one of the key items that affects the overall efficiency of defect detection is the false fail rate of a system. Simply put the more false fails you pass on to a repair operator for validation the more real defects get missed. A major study by Solectron presented at IPC showed that when false fails increased excessively, i.e. above 3000 ppmJ, an inexperienced operator could make <5% correct calls, while an experienced operators correct call rate still fell by over 50%.

Ensuring low false call rates can be a key aspect of a successful AOI implementation, increasing overall quality while reducing costs.

Figure 5 – The Benefits of low fails fail rates

♦ iv, What is your defect coverage, how do I get an accurate inspection coverage report.

Whereas many electrical test platforms such as ICT and Flying probers provide accurate reports on defect coverage most AOI platforms do not have the ability to provide this required feature. When you have an inspection routine in place how do you know it is actually finding defects? A few experienced vendors can provide this information, with the most capable AOI vendors providing validated defect coverage for all defects detected by the AOI system.

This is an important point that should not be ignored, their may be items on the product that may not be inspectable, knowing this will allow you to better plan your overall test and inspection strategy.

♦ 2. Improving your process and providing process feedback

Figure 6 – The Benefits of Parametric data from a rules based AOI system

Beyond catching defects what other aspects of an inspection system are important to manufacturing processes. Do you monitor how is your process changing? Can you predicatively highlight issues that will cause defects? If you are focused at improving your manufacturing processes, you will need variable data (measurements) vs. attribute data (pass/fail). If you are only looking at end of line defect data the attribute data produced by template/comparative systems may be adequate, if you are looking at understanding the process and correcting issues, access to parametric measurement data will allow for complete process control.

Figure 4 shows a simple example of a placement head going out of calibration. The top part of the graph shows the measurement information from a rules based AOI system and the control limits for the process. The lower part of the chart shows the attribute data from a non-measurement based AOI platform. The measurement system can provide data in real- time that can highlight the components placement going out of control far before it fails a pass fail limit is met. The benefits of process monitoring and SPC are well known and if these techniques are to be implemented in your process, the ability of the AOI system to measure placement and solder amounts will again increase the value of the AOI system.

♦ 3. Programming time

How fast is fast enough? This is a question the user of the AOI system will have to answer, but there are two factors to consider, first program development time vs. ongoing program support.

When having a demo of an AOI system most companies will have a program running in 1-2 hours or even less. However, check to see of this is realistic vs. how you will be using the system in your manufacturing process. Its important to look at programming (initial programming) and program support (program debug on new batches of products) times for the overall usage of the system and try not focus on the first demo program.

When using a template/comparative based system ask the question:

i. Is the program inspecting all defect types including solder?
ii. What is the false call rate, or on-going program debug required for the next batch of boards? How will this affect fault coverage?

When looking at rules based systems consider:

i, Is the system library based, if so how much does the use of the library decrease programming times over the lifespan of the AOI system?
ii, If we use rules based systems false fails will be lower, and will require less program support over time.

♦4. Stability, Support and MTBF

Will the AOI company be there to support you? Where does the company have support staff? Has the system been designed to provide high MTBF’s and reliability?

As recent events have shown us, large corporation have unhealthy focus PBIT (Profit Before Interest and Tax) numbers, plus they have high overheads and little focus on a single business. Even if a group is profitable, this has shown that it may not be safe. As such we have seen the exit of multiple major players from the AOI market that many would have described as stable entities that would not leave this business. In addition, some AOI companies have left the market a number of times, coming back under new ownership. This has left many customers with unsupported products with no path forward, and little value left in their investment.

There are a few companies whose sole businesses are focused on AOI. These companies have to providing best in class solutions to their customers; provide the best support, products, while keeping on the leading edge of technology. As AOI is their main and only focus, success in this industry is directly affect the success of their business. If the AOI company has this focus, has proven longevity and technology leadership, this will be a leading indicator that they are here to stay the course.

When it comes to reliability, MTBF is key. Look at the systems components, manufacturing practices used to build the system and the materials used. Does the system use a start stop motion technique or does the system using a linear motion system. For example a linear motion system will provide far greater reliability over time.

Applications and hardware support is important when running in any manufacturing environment. As your staff and personnel change, does the AOI vendor the resources to support you? Does the company have direct applications and hardware support engineers on a worldwide basis? Are those engineers located a reasonable distance from your facility and fully trained. These are important traits that vary greatly across many AOI manufacturers, and could affect your ability to be successful in the longer term.

♦ 5. Resolution and Throughput

Another common mistake made in an AOI evaluation is the focus on camera size, i.e. it is normally presumed that the larger the number on megapixels the camera has provides better performance. This is not always true; camera sensitivity (how fast you can take an image), resolution at the given FOV, what motion system is used (on a start stop system, is their extra time for the stage to stabilize before an image is captured), and the overall throughput at the same resolution are the aspects that need to be evaluated.

For example a system using a 1.4mp, high sensitivity camera, providing a resolution of 12um could have a throughput of 7sq in/s where a system with a low sensitivity 3mp camera only provides a resolution of 25um at 7sq/in per second. Design traits that will slow throughput include vibration from a high-speed move and stop, low sensitivity cameras, stage speed and overall FOV.

There are many factors that can affect throughput, but a defined resolution is required to detect defects, as a guide table 3 defines the typical resolutions required for inspection of the common component sizes. Understanding resolution vs. throughput will allow you to make a more informed AOI decision.

♦ Conclusion

Don’t make the common mistake and focus your evaluation on a single aspect of an AOI system. Megapixels alone are not important, programming times in a demonstration can be deceiving, and don’t lose focus on your real need for AOI.

Defect coverage, manufacturing strategy flexibility, the use of measurement data, throughput, company stability, system reliability, programming and on-going program support times all need to be judged equally in addition to the over all purchase cost of a system.

Defect coverage should always be the first reason for using any inspection platform. If you understand your manufacturing process and focus on the defects generated in your process the AOI system will always provide you with the maximum value and ROI, while increasing your overall quality.

by Paul Groome, Machine Vision Products, Inc. Published in US-Tech May 2009 Edition

In the current economic environment optimizing manufacturing costs, and especially Test and Inspection costs are high on most companies agendas. But we still need to ensure the highest level of quality for customer shipments. Both cost and quality goals can be achieved by carefully selecting the test and inspection strategy used in manufacturing. Understanding the benefits and defect coverage provided by the many solutions is key to achieving the highest quality at the lowest costs.

The majority of Test and Inspection techniques used today have been around for a long time. Most originated in the late 80’s when the prevalence of manufacturing processes were using through-hole technology. Why, with the introduction of SMT manufacturing processes over 15 years ago, are we still using the same equipment and test methodologies?

Most manufacturers I have worked with use the same techniques for Test as their companies used 20 years ago, ICT, Visual Inspection, MDA and Functional Test. Is this the most efficient way to ensure process quality and the lowest costs? No. In most cases Full AOI combined with Boundary scan will locate all production and electrical defects.

If we look at today’s manufacturing processes there are a plethora of automated test and inspection techniques to find defects and add quality; In-Circuit Test Systems (ICT), Manufacturing Defect Analyzers (MDA), Full Automated Optical Inspection Solutions (AOI-Full), Comparative Automated Optical Inspection (AOI-Comp), Functional Test (FT), Boundary Scan (BScan), Automated X-Ray Inspection (AXI) and Flying Probers (FP).

Figure 1 – Typical Package DPMO Rates

Each platform has capabilities that allow for different defect spectrum to be located, all at different cost points and diagnostic resolution. So, the question is which solutions should I use to meet today’s SMT Assembly challenges?

♦Defining a Test and Inspection Strategy for today’s manufacturing processes:

The first aspect of defining a test and inspection strategy is to understand the defects you are producing and what quality levels your customers demand.

The component types, board densities, and process equipment used drive the quality of your final assembly. DPMOJ rates (Defects Per Million Opportunity per Joint) for today’s packages range from below 50 DPMOJ to greater than 15,000 DPMOJ, with Area Array packages being the most reliable to place and reflow, while fine pitch leaded devices commonly having the highest DPMOJ rates. Figures 1 and 2 detail the typical DPMOJ rates experienced with customers at MVP (Machine Vision Products). If you are a member of iNEMI or IPC there are many tools available for you to calculate the capability of your process, expected yields and defect rates. For test and inspection it is important that your system capability matches your processes capability to deliver the highest yields.

♦System Capabilities: what is the cost to capture a defect?

Shown in figure 3, each of the test and inspection tools that can be deployed in a manufacturing process have different sets of capabilities and costs associated with them. Lets discuss the in-line options available to manufacturers today.

♦In-Circuit Test (ICT) and Manufacturing Defect Analyzers (MDA)

ICT system is the fact it can provide device function coverage for digital devices. Because of the time required to generate full test models most customers today are testing digital devices using capacitive opens techniques, taking the capabilities to a lead only test. One other issue with ICT is the fact it does not test parallel components, such as de-coupling capacitors and devices with multiple power and ground connections. Therefore fault coverage on a joint basis, commonly between 65-75%, a lot less than most customers realize. When loss of electrical access because of component density and signal frequency issues is taken into account the fault coverage falls further.

Today, dependant on configuration, systems ICT systems range from below $50,000 to over $500,000. When looking at ongoing support costs, the cost of a single fixture and program for a large board (>5,000 nets) can exceed the purchase price of an AOI (Automated Optical Inspection) system. The smallest boards can still cost $20,000, and take 2-3 weeks to complete.

Combine the cost of supporting ICT/MDA systems, joint based fault coverage, loss of access, with the fact that powered up testing for digital devices is being used less and less, does ICT provide cost effective fault coverage for today’s manufacturers?

♦Automated X-Ray Inspection (AXI)

Two major types of 3D in-line AXI systems are available in the market today. These systems are based on Tomosynthesis and Laminography. AXI can provide the highest level of lead based inspection coverage, but have a number of issues associated with providing this coverage. Normally a 3D AXI platform will range between $450-750K, in most cases be unable to meet line cycle times and can provide false calls above the 5000 ppmJ range.

Programming can be complicated and slow, as the varying track densities of the PCB substrate can affect each joint causing the programmer to treat many joints as separate items (Sub-Joints).

Today unless you are specifically looking very explicit solder defects on very high cost products, it is not a viable option for most manufacturing processes because of costs and throughput.

♦Boundary Scan (BScan)

As part of the Joint Test Action Group (JTAG) boundary scan became a standard in the 1990s, showing great promise. But the adoption of boundary scan was slow until recent days where loss of electrical access has become a large issue with ICT and MDA systems. Boundary Scan, when implemented, provides the highest level of digital fault coverage at the lowest cost per defect, with solutions ranging between $10,000 and $25,000.

The key to Boundary Scan is to ensure your circuit design engineers take into account the capability. Most larger scale integrated circuits today include Boundary Scan. If your design engineers chain these devices together correctly, Boundary Scan techniques can provide very high fault coverage. Boundary Scan is one of the best tools in the industry today for digital device coverage, ISP and flash programming.

♦Automated Optical Inspection (AOI) – Full AOI and Comparative AOI

Most people bound the two different types of AOI (Comparative and Full Rules Based) systems into one category, but these two types of systems have substantially different capabilities and performance in the manufacturing process.

When new users start to look at AOI, in general they start evaluating aspects that are easy to understand like: megapixels, magnification, optics, all of this is nice to know but does it mean anything. Is a system with a 5 megapixel camera better than a system with a 4 megapixel camera? As single items, the specifications of each of the components are insignificant. The real specifications to evaluate are how the components work together which in-turn drives the pixel resolutions vs. the throughput of the systems. For 0201 style components a resolution of 16-18um will be adequate. Using a real-world example, a system may quote a 4 megapixel camera, have throughput of 5 Sq inches per second and pixel resolution of 25um while a second system could use a 5 megapixel camera provide a 7 Sq Inches per second throughput and have pixel resolution below 16um. Clearly when looking for resolution and imaging capabilities the overall system capabilities need careful critique.

♦Full AOI

Full AOI will provide you with the best defect coverage, highest throughputs, repeatability and metrology. The main reason for the increased performance is that these types of systems actually measure the parameters of a component and joint being inspected rather than comparing to a suggested good part. As such these techniques are less susceptible to process variations while providing enhanced inspection techniques to measure components and parameters of the product. Unlike Comparative AOI systems, Full AOI metrology based techniques can provide extended capabilities that can allow for full SPC, process control and component measurements.

“The techniques used by full metrology based AOI systems are far less susceptible to process variations while providing enhanced defect coverage over comparative based AOI systems.”

When choosing a Full AOI system, there are some key aspects to look at, one being the techniques used to locate defects. The most powerful of these techniques is the Tri-Color technique that allows for simple, accurate and reliable defect detection. This works by using angled light vs. angled cameras to generate solder and joint information. For example three rings of Red, Green and Blue LEDs are used at different angles (0o, 45 o and 90 o) each showing different aspects of the component, joint and solder being inspected. Figure 4 shows an example of a lifted and acceptable QFP joints that can be easily detected using these techniques.

Figure 4 – Lifted QFP cornet pin detected using Tri-Color Technology

The second point to look at is software. Many Full/Metrology based AOI systems can be time consuming to program for the first article. This is not true for all systems. Companies like Machine Vision Products have been investing heavily to remove the gap between Full AOI and Comparative AOI programming times with products such a ePro that fully automate the program generation process effectively providing you with the fastest programming times without the loss of coverage and performance you would normally see with comparative based systems.

Full/Metrology based AOI systems also lend themselves to other micro-electronics and Semiconductor environments where understanding the process and the provision of SPC data is of the utmost importance.

♦Comparative AOI

These types of systems use a golden image of a component and compare the part to the component under inspection to provide Pass/Fail information. Their focus is predominantly on placement, presence/absence, and optical character recognition. These techniques can provide a quicker initial program, but has the issue of decreasing performance in production, as variations in your process have to be taught as you see them in production. Creating high false fail rates per board. These types of systems can generate programs very quickly for first article inspection but provide significantly reduced performance and risk in production.

♦Choosing the Correct Test and Inspection Strategy

Figure 5 details the Total Defect Coverage for each type of test and inspection scenario. As discussed previously, all have different capabilities and attributes, but in today’s environment where quality and cost is key you need to choose the best solutions and most cost effective Test and Inspection equipment to meet your manufacturing needs. Full AOI clearly allows for the maximum fault coverage at the lowest cost and can be complemented by a number of different solutions.

Figure 5 – fault Coverage Capabilities, NPI and Lowest Cost Strategies

For example, if you have designed in Boundary Scan, an unsurpassed level of defect coverage can be obtained by distributing test and inspection with Full AOI. Figure 5 shows the defect coverage capability. On going costs, when distributing test and inspection across Full AOI and Boundary Scan are extremely low, minimal fixturing is required, and programming time is commonly less than a half a day. Using Full AOI also allows you to use deploy process control and using these two platforms. The deployment of Full AOI and Boundary Scan in a distributed test and inspection strategy can provide: 100% component coverage, determine the values of passives, ensure the operation and functionality of each digital component on a board while costing significantly less than any of the traditional test methods used today.

Many combinations can be used, but the obvious winner for cost reduction while still providing the ability for total defect coverage is a distributed test and inspection strategy that includes Full AOI. The Full AOI system needs to incorporate: metrology, tri-color defect detection techniques, flexible measurement capabilities, high resolution and ease-of-use programming features combined with a combination of Boundary Scan or MDA platforms.

by Dr. George T. Ayoub, President & CEO, MVP Inc.

High speed AOI inspection solutions have been well received for over than 2 decades by several industry sectors such as traditional SMT, automotive, and PCB manufacturing. The innovative solutions, built in quality, reliability, low maintenance, and state of the art high speed inspection capabilities have attracted several industry leaders and High Volume assembly powerhouses around the world.

Last year MVP introduced yet another innovative line of products which has revolutionized the AOI technology in several electronics industry sectors. Targeted towards packaging applications, the new platform was designed to meet and exceed today’s complex High Volume Manufacturing assembly requirements. The configurable nature of the platform makes it a perfect high speed AOI choice for many applications in complex and hybrid C4 + SMT assembly lines.

♦A New Inspection Platform for Packaging Applications

Market demand for new electronics products has been on a fast track with growing desire for slick features, mobility, and increased integrated functionality. Aggressive design requirements mandate smaller form factors forcing shrink in all three dimensions. Integrated functionality, on the other hand, forces higher mix of SMT components with smaller features and lower profile. This trend provides a significant challenge for product integration specifically in the area of packaging assembly.

Real time inspection with right ‘’hooks’’ in place to provide meaningful feedback and easy to digest output information can now enabled complex assembly lines to become more efficient in managing upstream-downstream process in terms of line yield, utilization, overall productivity, and profitability. The platform was designed with all of the above considerations in mind to meet today and tomorrow’s AOI needs.

A configurable platform, dependant on the application, uses different electro-optic and-or Material Handling Systems can be combined to meet various processing requirements. However, the base platform is the same for all applications. The similarity between different configurations improves overall utilization of complex assembly lines. Once basic training has been provided, equipment operators can switch positions from one process to another with minimal training since the same operating system and interface is shared across all configurations. Other key advantages are spares management and equipment maintenance. A very large portion of the same components is shared among all configurations which ultimately helps with spares management by reducing the number of spare parts on the shelves, enhances troubleshooting, and periodic maintenance.

All configurations are equipped with a solid granite stage to enhance overall inspection precision. A high precision frame to complement the granite stage, a single 4 mega pixel color camera, and programmable LED lighting enables repeatable high speed on-the-fly image acquisition with field of view resolution of 3-25um/pixel. Telecentric lens is an option to further increase inspection accuracy required for certain applications.

Careful attention has been given to the Material Handling System to assure ultimate flexibility and meet JEDEC tray standards, metal carriers, bare PCB as well as thin strip processing. All of the platforms can be configured as single or dual lane. Support pedestals and auto board clamping are also optional for more precise board registration and handling. Upstream-downstream communications have been taken into account with flexible PLC control, SMEMA interface, and full automation capability to transfer tool performance and recipe specific data to line management servers.

The platform comes with a powerful integrated SPC package. A large stream of valuable inspection results is gathered continuously and can be plotted real time in different formats to assist troubleshooting and maintain a high yielding assembly line. Offline programming and debug is available to minimize production interruption. A CAD driven library based programming Software cuts new recipe creation and testing time.

With the ever growing evolution of packaging technology, specifically in the organic packaging, and thin package processing the need for AOI continues to increase. The MVP 850G platform was specifically designed to address all new and next generation packaging assembly needs. The tool can be configured to perform 3D Paste Inspection, 2D Flux Inspection(without florescent additives), C4 Die and SMT component Inspection(pre and post reflow simultaneously), C4 Epoxy Underfill (spread, quality, fillet, etc), Surface Finish(scratch, damage, etc), Wire bond, Glue and Sealant, traditional SMT (pre and post reflow), and many more applications. The tool can be placed in-line or off-line depending on assembly layout and process needs.

♦Die Placement Metrology System

MVP were given a challenge by one of their major customers to provide a solution for the accurate measurement and inspection of dies placed onto a substrate. As the placement of these dies are critical to the reliability of the products MVP engineering and management embarked on a project to develop a measurement and inspection tool.

The tool required to be a robust metrology based system to be capable of a repeatability for translation in X and Y of 1.3 microns, reproducibility in X and Y of less than 2 microns, die rotation repeatability and reproducibility not to exceed 0.007 degrees and overall accuracy between different tools not to exceed 10 microns in total.

With years of experience of a variety of different inspection approaches across many industry sectors a taskforce set about specifying a new inspection tool which would meet the requirements of the customer. Not only would the tool have to be capable of inspecting the die placement on the substrate both post and pre-reflow, it would also have to be capable of the inspection of surface finish including scratches and irregularities on the surface of the die.

Other capabilities would include the inspection of surface mount components such as 0204, 0201, 0603 IDC, 0402, resistor networks and even 12 mil pitch QFPs.

In order to meet the variety of inspection requirements a proprietary electro-optics module was developed. Resolution studies showed that a 16um pixel size is adequate for meeting the various requirements of speed and accuracy. The electro-optics module utilizes a telecentric lens, and a tri-color lighting source complemented by a white light source. This assured visibility to the surface mount defects as well as an improved signal to noise ratio for the edges.
The platform software is equipped with a range of existing inspection algorithms including sub-pixel edge detection, surface defect detection, post and pre-reflow SMT and metrology algorithms.

Other significant challenges were to meet the number of units per hour to be inspected. In the range of 3000-4000 UPH required the utilization of concurrent fiducial registration and inspection which allowed for significant time savings during the inspection cycle.

The system required to be provided with a dual lane capability, this effectively working in tandem with existing production equipment to maintain the UPH of the line. Additional challenges were the requirements to integrate SECS/GEM and Lot Code Traceability to the inspection tool.

Another critical parameter was the uptime of the systems which had a base specification of greater than 98.5% production availability. Again the platform surpassed this requirement providing greater than 99% uptime.

In the following charts it can be seen how the platform not only met the customer specification but significantly exceeded the specification.

♦Repeatability and Reproducibility Sigmas

♦Matching Test with Benchmark System

The following chart shows results from one of the metrology based systems. Sixteen dies with various offset (DX) were measured by a benchmark systems and the die placement metrology system. Note that the slope is 1.032 from the linear regression fit and R-square is 0.99. Bias is 1.4147 microns.

Not only were MVP successful in the implementation of the first tool to the customer, they have since provided over 50 similar systems to the same customer which all have passed rigorous test routines prior to acceptance.
With the transition from traditional SMT inspection into micro-electronic inspection capabilities MVP have proven their strength lies with high performance, flexible and innovative inspection solutions.

As Seen in Advanced Packaging, September 2004 by George T. Ayoub

Flux inspection has been a challenge for Flip Chip and BGA assemblers due to the inability of inspection systems, including AOI (Automated Optical Inspection), to accurately see the material and therefore be able to inspect it while maintaining line speed. Specifically with respect to flip chips, the inspection of flux is an important part of controlling the process and can prevent costly mistakes from happening. Fortunately, a machine vision solution utilizing UV (Ultra Violet) illumination, which has been developed over many years of research, is able to detect defects from flux deposits. This technique, which has now proven successful for more than three years on high volume manufacturing lines, replaces the visible AOI light with specialized UV light that matches the properties of the substrate and flux, in order to achieve optimized inspection results.

♦Importance of Inspecting Flux in the BGA/CSP Assembly

Flux plays a critical part in the process dynamics of BGA/CSP package assembly. A vast range of defects in the final assembly can be traced back to poor flux or paste deposition. For example, some of the defects in the final assembly derive from poor flux alignment with respect to the intended pads, insufficient thickness/amount of the flux material, excessive amount of flux, or from smearing. The detection of these pass/fail types of defects (attribute data) at an early stage of the process reduces the assembly cost significantly. Moreover, many manufacturers would agree that it is important to control the process of flux deposition by means of relevant measured variables in order to detect trends and prevent defects from happening in the first place. This requires a system that is able to measure the key variables of the process (variable data). By providing real-time information on key process parameters, manufacturers can take corrective action and prevent scrap and production loss.

♦Technology Challenges: Making the invisible visible.

The process of flux inspection flux has been a challenge to AOI manufacturers due to the inability of visible light to image well the flux material and therefore to be able to inspect it. Both high and low angle illuminations in the visible suffer from poor signal to noise ratio of the flux with respect to the background. However when the flux is illuminated with UV lighting, it fluoresces in the visible and the signal can be captured by means of proper filters designed to eliminate any background light not emanating from the fluorescent flux. Under these conditions the signal to noise ratio between the flux and the background is enhanced significantly. The key to obtaining a good signal to noise ratio is the proper design of filters and illumination -which are proprietary – that are adaptable to the flux itself and to the background material (ceramic and possibly FR4) and at the same time able to eliminate the visible background light. (See Exhibit “A” – Images under visible light versus UV light.)

♦Challenge: Speed and resolution for an in-line system.

Image acquisition parameters play an important part in the capability of the system. Two important parameters are the speed of inspection and the proper optical magnification (resolution). Both are related because speed of acquisition is inversely proportional to the number of pixels acquired, which in turn varies linearly with the square of the magnification. Moreover, illuminating small areas require a large amount of light and an adequate number of pixels on target. The requirements to keep up with very fast cycle times coupled with the high resolution were met by means of utilizing multiple cameras heads, proper illumination utilizing UV diodes and specialized electronics. Multiple camera heads (in this case three are used) extend the Field of View from square to rectangular shape, while at the same time not sacrificing the resolution. The specialized electronics allows the camera to acquire in parallel and matches their speed to the processor computing speed. Utilizing UV diodes assures longevity and stability of the system over time, which is extremely important when an inspection program need to run without modification on different systems or production lines.

The system is also capable of measuring the paste parameters under UV and/or regular visible light making it extremely useful for controlling the paste deposition process at the same time. The algorithms, that extracts position, area, pad area coverage and brightness are based on blob analysis and are only applied in the areas of interest.

♦Defect detection, measurements variables and SPC

Utilizing UV fluorescence techniques, the system in operation goes beyond the detection of the pass/fail defects attributes, and assists in enhancing yield by means of SPC techniques on measured variables. It measures the position, area, pad area coverage, and brightness of the deposited flux. Brightness is measured by calculating the median gray scale of the paste blob. Although it is ideal to utilize the height and volume of the flux, these measurements with the UV technique utilized here, are seen to be dependent on the properties of the flux and the background where the flux is being deposited, and cannot be trusted in all cases as absolute measurements. There are good logical reasons backed by experiments confirming that brightness correlates with the height of the flux. In effect, the brightness depends on the amount of fluorescent material in the flux and therefore should vary linearly with the volume. However this linearity is not always certain, but depends on the environment. Therefore, care should be taken when interpreting the measured brightness since other materials may fluoresce also and add to the noise. The method described has proven to be effective in the production environment, utilizing brightness along with position, area, and pad area coverage as measurement parameters, thus providing a logical and adequate means for controlling the final quality of the process by means of SPC methods.

In the production environment, real time process control has proven to add value to the process by following trends and preventing defects from happening, (See Exhibit “B” – Real-time data charts) and has been an integral and critical part of the system. Depending on the alarm setting, the system is able to stop the line or turn on a yellow or red light for visual feedback to the operator.

♦Conclusion

The described technique has been proven for over three years of inline inspection. The system is process capable with GRR in the range of 2.5% to 8. It is able to keep up with a relatively fast production line speed while achieving a false call rate in the range of 10 to 20 ppm and a false accept rate less than few ppm. By containing defects at this early stage and by controlling the trends with SPC, good results have been achieved. Future work planned is to keep enhancing the signal to noise ratio and to extend the application of this technique to different substrates and flux types.

References
Reliability and Yield in Flip-Chip Packaging, Alan Lewis, Ed Caracappa, Lawrence Kessler, 1998_11_hdi_flip_chip_reliability.pdf