Connecting leads to die involves precise tool selection.

Wire bonding is a process that creates an electrical connection between a die and a substrate or lead, typically using gold or aluminum wire. Wedge bonding is a specific type of wire bonding that uses a wedge-shaped tool to create the welds. The design of the wedge tool has changed very little over the past decade. The wire is fed at an angle through the back of the wedge. This angle is typically 30° to 60° and is application-dependent. Some applications require a higher feed angle because of package clearance issues. Some deep access applications require a 90° feed angle. In this configuration, the wire is fed (Figure 1) through a hole in the shank of the wedge tool.

A wedge has a number of features that contribute to the quality and consistency of the wire bonds and loops. A thorough understanding of the wedge geometry is required for selecting a suitable wedge for any application. Figure 2 shows a detailed view of some of the critical features of a wedge-bonding tool.

Bond length is calculated using the formula:

BL = 2/3FR + BF + 2/3BR

where
   FR = front radius
   BR = back radius
   W = width of the wedge
   ϴH = hole diameter
   H = wire feed angle (30° through 90°)
   BF = bond foot length
   BL = bond length
   VR = vertical relief
   T = total length of the wedge

Special considerations should be taken before selecting a wedge tool for an application. Here are some of the critical design features that drive wedge selection.

Wedge material. The wire material plays an important role in determining the wedge. A tungsten carbide wedge is used for aluminum wire, and a titanium carbide wedge is used for gold wire. The wrong material will cause premature wear of the wedge tool, resulting in poor welds. The latest developments in bonding tools for fine-pitch applications focus on implementation of new materials such as cermets, void-free carbides, and ceramics to increase long-term stability and to avoid interactions with wire and pad materials.

Bond pad pitch. Pad size and pad pitch play a significant role in selecting the wedge. Typically 100% of the bond is required to be on the pad. Bond width, bond length and pad pitch should be considered before selecting the wedge. (Bond pad pitch less than 50 µm is considered fine pitch.)

Wire feed angle. Loop consistency is critical, especially in stacked die applications with inner and outer loops. Lower wire feed angles give better loop control, while higher feed angles give less loop consistency. Figure 3 shows the importance of wire looping to avoid shorting. Higher wire feed angles are preferred for deep access application.

Back radius. For applications where tail consistency is critical, a sharper back radius is required. However, sharper back radii can result in heel cracking. A tradeoff in the back radius is critical for achieving uniform tailing and avoiding heel stress after the first bond. A normal back radius will range between 50 to 100% of the wire diameter.

Front radius. The front radius and the bond length chiefly define the strength of the second bond. The front radius ensures a gentle transition from the wire into the weldment of the second bond.

Wire hole diameter. The wire hole diameter plays an important role in applications that require high placement accuracy. A typical guide for selecting the hole diameter is two times the wire diameter. If the hole is too large, the wire will tend to move under the foot of the wedge. This will result in poor accuracy (and poor looping). If the wire hole diameter is too small, the wire can scrape along the sides of the hole as the wire is fed through the wedge. This will result in wire slivers that can cause shorts and reduce reliability. Wire slivers can be prevented by requesting wedges with polished holes to remove any metal burrs.

Vertical radius. Vertical radius is also referred to as vertical side relief (VSR). High-frequency devices can have large numbers of I/O pads placed as close to each other as possible, or very small individual bond pads. Wedge bonding can achieve finer pad pitch geometries than ball bonding with the same wire diameter because of the smaller amount of bond “squash” or wire deformation. Wedge bonding squash is typically 1.3 times the wire diameter as compared to two times the wire diameter for ball bonding. A thinner vertical radius permits wedge bonding to smaller bond pads without the need to decrease wire diameter.

Tool face geometry. Cross groove tools increase the ultrasonic coupling and are recommended for hard gold wire bonding. Soft materials like aluminum wire are not recommended for use with a cross groove tool because of the possibility of aluminum buildup in the groove. Flat or concave tools usually are used when bonding with aluminum wire.

The American Competitiveness Institute (aciusa.org) is a scientific research corporation dedicated to the advancement of electronics manufacturing processes and materials for the Department of Defense and industry. This column appears monthly.

A series of companies describe how they use inspection.

Defect prevention and containment is a very important part of today’s production. Cost pressures are greater than ever and quality is scrutinized in new ways. This, coupled with the speed at which new products are being generated, and the advancements in technology with finer pitches and more detailed components, put PCB assemblers in a challenging position. The choices and options for automated optical inspection are many. AOI can be implemented in multiple ways if you consider the location (i.e., pre- or post-reflow) along with inline vs. offline inspection. It is difficult to navigate the murky waters to determine the AOI test strategy implementation best suited for a given product, production environment, and to yield the best results. This is particularly visible in lower-volume environments with a high mix of products. In high-volume manufacturing, most lines are equipped with AOI. In many cases, they have 100% inline post-reflow, solder paste inspection and in some cases, pre-reflow inspection too. But for a situation of high-mix, low-volume, or just a few lines, it becomes complex to balance the investment and return for AOI. I recently sat down with several AOI users from different product areas and assembly segments to see what they are doing. A few of their summaries are described, along with some conclusions from the exercise. This may provide some considerations to help with the implementation considerations facing your organization.

Low-volume, high-mix telecom. One telecommunications company building products in a low-volume, high-mix environment uses AOI in a pre-reflow function. It has been doing so for about 18 months. Every board is inspected at 100%, through inline pre-reflow AOI. The company mainly catches missing parts, misalignments and polarity issues. The final yields did increase within the first few weeks/months after implementation of AOI in the pre-reflow position, and there has been a consistent stream of caught defects over time. Funding has prohibited the adoption of SPI and post-reflow inspection, but those are desired. On a daily and weekly basis, the AOI data are compared to the automated x-ray inspection data to address major issues and quickly localize the area of production impacting the issue.

Defense-based OEM. An OEM building medium-complexity applications for military use has two surface mount lines. The manufacturing environment is low-volume, high-mix. It uses one offline AOI system, colocated with production, to service both lines. Its AOI is used for post-reflow and has been in use for about seven years. The motivation for post-reflow in this instance was to combat the investment in in-circuit test fixtures in a high-mix environment; over seven years, the company has relied more each year on AOI. For newer products, the main failure observed at AOI is wrong parts or incorrect polarity. For more mature products, tombstoning or missing solder have been the most prevalent defects. Results are pulled daily, and alerts sent to the various product steps, AOI programmer, process engineer, etc.

EMS in medical segment. A contract manufacturer focusing mostly on medical had been relying on manual visual inspection for its low-volume, high-mix environment. Internal studies show its MVI process is about 50% effective. With continued pressure to improve quality but balance cost, the organization completed an AOI analysis. The fast inspection times, inline or batch capabilities, and repeatability clarified the conversion benefits very quickly from both perspectives. The polarity, presence and absence, and part recognition features solidified the AOI benefit. The AOI systems are used post-reflow and are positioned inline and have been in use for about two years. In this company’s case, constant data scrutiny and correlations to the AXI data have contributed greatly to the gains from the AOI implementation.

The high-mix environment lends itself well to AOI utilization thanks to its fast and easy programming. The improved quality witnessed quickly returns on the initial capital investment, although most will not share data in detail due to their proprietary nature. In most cases, post-reflow is the preferred inspection location, although for telecom it’s very common to see pre-reflow or a combination of pre- and post-reflow. The inline positioning in general seems preferred, but offline implementations are seen in cases of very low volume or situations where funding for systems to support all lines is not available. In general, those who use AOI recommend AOI.

Stacy Kalisz Johnson is Americas marketing development manager at Agilent (agilent.com); stacy_johnson@agilent.com. 

On the permanent hunting for efficiency.

Many believe in the theory of evolution, many do not. What you believe does not detract from the situation facing us all today. Either belief can accommodate the notion of survival of the fittest, and the economic crisis raises the bar to the point where only the fittest company, organization or individual makes it to the next period of economic prosperity.
During financial booms, very few are concerned about its inevitable end, and the system copes with a huge range of business acumen and abilities. Efficiency becomes less important. Individuals or organizations that are not particularly efficient can flourish and not worry about the future until a recession comes along and then – shock, horror! – workers are made “redundant.”

This is sad, of course – and can be catastrophic for those directly affected. We have two major events to consider from this: the need to smooth out the peaks and troughs of economy, and the effect on supply chains. Peaks and troughs have always existed. Time was, the machine tool industry assumed a five-year cycle between the highest peak and the lowest trough, and similar theories exist today, but there is much more reliance on the supply chain feeding itself properly now.
Because we have insisted, quite rightly, on becoming efficient and use techniques such as TQM or JIT, we have developed a supply chain wholly dependent on itself for survival. In the days when we used to build stock and there were economic cushions, the supply chain could survive relatively long in a recession and the effects were less dramatic and more drawn out. If the likes of a GM or a Ford catches a cold today, everyone sneezes tomorrow rather than weeks later, as would have been the case 20 years ago. There is no fat left to live on when times get hard.

Therefore, when people stop buying cars, not only do the car manufacturers suffer, the suppliers suffer as well – and immediately. In the UK, Honda decided to cease making cars for two months because of falling sales. Their suppliers, in turn, suffer immediately. The electronics module suppliers cannot supply; their assemblers cannot supply; the component suppliers cannot supply, and then the PCB manufacturers and the laminate suppliers cannot supply. Where does it end?
The recession is here to stay for a lot longer. There will be many company failures. The point is will we lose so many suppliers that we cannot recover? Every day, we read that company A is restructuring and cutting costs, and in many cases, they close. If we kill off so many important suppliers in the chain, then we will never recover. Recessions in the past affected everyone badly for a while but, in the end, we all got through. This time, it may be different.

One hundred-fifty-years ago, there was a nasty event called the “South Sea Bubble” when large-scale investments were made in schemes that were not substantial. The real losers were not the investors, but the subsidiary companies that were sucked into the mess. Is history repeating? Have we not learned? We have learned how to become efficient, but not how to preserve the supply chain. We must now rely on the survival of the fittest if any suppliers are to stay with us.
Unfortunately though, we still allow various managers and supposed entrepreneurs to mismanage companies and institutions and walk away with no penalty. Indeed, in many cases we reward failure. How can this possibly equate to the survival of the fittest? The obvious conclusion is that such mismanagement comes from unfit people, and we should immediately fire them with no recourse, instead of letting them sack thousands in return for a pay raise.

Some of the mismanagers take risks without worrying about the consequences because they are safe in the knowledge that some corporation or government will bail them out. Bailing out a large manufacturer such as a car manufacturer is acceptable, provided the reasons for the car manufacturer’s problems are not associated with bad management, but rather unfortunate timing and a public short on funds. In this case, there is a chance the supply chain can be maintained. Bailing out a failed bank is a different kettle of fish. It is very possible the bank’s managerial team “played” with its client’s money without worrying about the consequences of its actions, safe in the knowledge a golden parachute would cushion any bad decisions.

Fitness now includes the need to monitor trends and predict probable cyclical ups and downs so as to ensure the possibility of maintaining a supply chain. We must be aware that the supply chain has no excess fat and that we have driven it to be lean, mean and fit. Unfit managers can no longer be tolerated. Fitness for purpose includes every single member of a company working to their maximum efficiency all the time with no safety net for failure. This principle applies to public corporations as much as private companies. Public corporations must not expect floods of government money if they encounter trouble; their managers must be as adept at maintaining efficiency as their private industry counterparts.
If we want lean, mean and fit companies, we must be permanently hunting for efficiency drives. There might then be many more of us who can be considered “fit for purpose,” and the effects of recession should not be so severe.

Peter Grundy is director of P G Engineering (Sussex) Ltd. and ITM Consulting (itmconsulting.org); peter.grundy2@
btinternet.com. His column appears bimonthly.

Designing in flexibility isn’t easy, but it’s a must.

The “Have it Your Way” slogan popularized nearly 25 years ago by Burger King defined the requirements for the fast food industry. Understanding that customers want choice, flexibility and selection to satisfy their unique and individual tastes set the tone for the future of the fast food market, and those who couldn’t offer variety likely didn’t survive long.

The “customer is king” mindset has transcended many other industries as well, and now, as we find ourselves in the midst of global financial chaos, this requirement is even more profound. Customers want choice. Customers want the ability to buy for today’s needs with the adaptability to modify for tomorrow’s requirements. Customers want value, of which flexibility is an inherent attribute. This is certainly not a new philosophy, but one that is now more poignant as we continue to navigate the current recession. When it comes to equipment systems, designing in flexibility from the foundation isn’t always the simplest route, but for customers, the ability to choose is essential to lowering cost of ownership and planning for future needs.

The platform equipment concept certainly isn’t new to electronics assembly. For years, notable names in the pick-and-place segment have offered limited scalability options to enhance or upgrade equipment performance. For example, one popular placement machine offered the ability to select one or two placement heads plus tray feeders as performance options. Historically, delivering this type of flexibility for the screen-printing market, while available, hasn’t been as customer-friendly as it should be. Sure, you could spec in options from the outset and a machine could be built as specified, but adding upgrades at a later date was limited to a few choices, and incorporating these technologies after the fact wasn’t always a seamless process.

But that is changing. What customers want with all of their systems is the ability to add options – numerous options – when and if they want or need them, and now they can do just that. As I stated, designing and engineering a screen-printing platform so that it can be retrofitted in the field is quite a difficult proposition. In fact, from an equipment design point of view, it means adding constraints to the project that elevate the engineering complexity and design time. Clearly, creating true modularity isn’t the path of least resistance, but it is arguably the best solution for the customer.

With the platform concept, consider that every single retrofittable option has to be designed to be compatible throughout an entire equipment range. Not only does each technology have to be mechanically harmonious, it has to fit within the cover packages, the software, the electrics; a whole host of extra work must be done on the front-end to achieve meaningful modularity in the field. From an equipment manufacturing perspective, it also means the process must be very lean, agile and flexible because, in theory, every machine going down the line potentially could be unique.

With advances in software and equipment design techniques, suppliers have made large strides in delivering a truly on-demand, customizable platform. Customers can now decide where they want to fit within the product curve today in terms of cost of ownership, flexibility and output. And, they can then modify their criteria when production volumes or conditions dictate a change. A printing system is selected based on process alignment capability and throughput as a starting point. From there, virtually any option can be added to any of the platform systems to address manufacturing requirements or preferences. For example, say you are a speed animal and your biggest concern is throughput. Obviously, you select the system with the fastest core speed, but you can also add a super-quick understencil cleaning technology that enhances throughput even more. Or, suppose flexibility is the most pressing concern for your operation. Now, with new system modularity, you can add in – at anytime, not just with the PO – different cameras for alignment or a bigger field of view. And, of course, if throughput and flexibility are equally important, customers can retrofit tools to address both requirements.

At Apex this year, for example, we saw most new printers now have the option to provide at least limited dispensing capability, either through direct or third-party add-ons. Outside the printing spectrum, we even saw a machine that contains screen-printing, solder paste inspection, adhesive dispense, placement and post-placement inspection all in a single platform. A true one-man band.

From board support, understencil cleaning, inspection, print verification, dispensing capability and more – just about anything the customer wants, the customer can have when and how they want it. The customer is, indeed, king. Choice and modularity rule and you can “Have it Your Way.”

Clive Ashmore is global applied process engineering manager at DEK (dek.com); cashmore@dek.com. His column appears bimonthly.