Market forces continuously drive test-handler companies to improve their products. Fortunately as these market forces unfold so do many technological advances that make the development of new test-handling methodologies possible.
Advances in packaging technology and the growth of new package types such as ball grid arrays (BGAs) and chip scale packages continue to challenge handler manufacturers. The 1980s saw the beginning of the proliferation of pick-and-place handlers. This was prompted by the popularity of quad flat pack (QFP) devices that could not be handled in gravity-fed machines, the most widespread type of machines in use at the time.
Current advances in semiconductor package design result in devices that are thinner, smaller and have finer pitches—if they have leads at all. Many of these parts present major hurdles to traditional pick-and-place handling equipment which already uses hardware-based conversion kits to accommodate different package configurations.
Minor variations in package design often require separate kits, resulting in large capital expenditures and lost time waiting for new kit design and fabrication. These problems call for advances in test-handling methodology beyond pick-and-place.
Higher Performance Testing
Today, it is normal for test requirements to be several gigahertz. As a result, traditional handling and contacting techniques are pushed to the limit when testing at these frequencies.
Many techniques can be used to design contactors to operate at high frequencies. The easiest design rule keeps the distance from the device under test (DUT) to the tester electronics as short as possible. Consequently, handlers should plunge-to-board or present the DUT directly to the test-head electronics rather than to a contactor where the device could be several inches away from the electronics.
Improved Semiconductor Yields
Until recently, capital investments for semiconductor manufacturers primarily focused on improving yields, or the percentage of good devices in a batch or lot, rather than on maximizing productivity, or how many good devices the handler/tester system can process in a given unit of time. Today, this is shifting dramatically.
Figure 1 illustrates the reason for this shift. Fabrication yields have improved over the past few years. As this happens, the return on investment (ROI) increases by investing in productivity improvement rather than in further yield improvement.
Rising Cost of Testers
Today, a test system could cost upwards of $5 million. Although the test handler may cost less than 10% of this amount, it is the handler that determines how much the tester will be used. Expressed differently, if a handler could offer twice the productivity, then only half the number of $5 million testers would be needed. Therefore, the pressure to improve the productivity of the handler goes up as the cost of testers increases.
There is a payback: Although the cost of testersis rising, their capabilities continue to grow. Enhancements such as higher clock speeds, more pin electronics cards and flexible architectures contribute toward increasing tester and handler productivity, such as parallel testing.
No longer the sole domain of memory test, parallel testing of nonmemory devices is making inroads in mainstream production testing. Parallel testing results in much higher use of ATE and, consequently, greater handler productivity.
Advances in Vision Technology
Vision technology has progressed rapidly over the last few years. For some time, cameras have been integrated into industrial equipment applications such as circuit-board assembly and placement equipment. Now, machine vision is finding its way into new test-handler designs that provide many features and benefits. These include:
Elimination of most expensive and complex conversion kits required in traditional pick-and-place handlers while allowing a large variety of package types to be handled in a single type of machine.
Vision that helps to reliably contact devices with extremely fine lead pitches.
Cameras built into the handler to integrate lead inspection into the test handler and possibly eliminate the need for a costly process step.
The fact that a number of well-capitalized and financially stable test-handler companies exist today is one of the most important changes to this segment of the semiconductor capital equipment industry. These companies will develop new test-handling solutions that might not have been possible as few as even five years ago.
Improved yields are driving the need for greater productivity in test handlers. In response to this, expect to see a great deal of emphasis on designing in productivity improvements in upcoming handler announcements. For example, machines will be faster, with higher throughputs and shorter index times (the time from end of test of one device to start test of the subsequent device, or the idle time of the tester).
Devices will be tested with higher levels of parallelism. And automation such as automatic tray stackers and tube-handling mechanisms will be regular options on new designs.
Finally, the drive toward process integration is forcing handler companies to think how their machines can combine with machines doing other back-end processes. This could save process steps and associated costs such as labor, space and in-process inventory.
The proliferation of device and package types is driving handler companies to maximize the degree of flexibility designed into their machines. This flexibility is extending beyond today’s common strategy of designing machines that can be reconfigured to handle a variety of device types by using conversion kits that are, often times, complex and expensive.
In the future, expect to see approaches to machine design that require little or no conversion-kit parts. This will be accomplished while maintaining or even increasing the degree of flexibility to handle a variety of package types. Along with package flexibility, new handlers will interface to test contactors manufactured by virtually anyone the customer wants.
The useof machine vision for accurately locating and precising devices is one of the most important technological advances in test-handler design of the last few years. In the near future, expect to see this feature on more advanced handlers.
New Moving Techniques
Smaller and thinner devices, finer pitches and entirely new packaging techniques are prompting handler designers to explore methodologies beyond gravity-fed and pick-and-place. Among the more promising is linear motion technology.
Linear motion technology consists primarily of precision belts that transport naked devices such as QFPs and BGAs through the handler without the need for the pockets, boats or carriers common in today’s pick-and-place handlers. This increases handler flexibility and productivity while nearly eliminating conversion kits.
Software tools in widespread use today, such as Windows NT, offer advances in system controllers over what was available for handler designers just a few years ago. These tools, in conjunction with ever-increasing computer power, also improve operator interfaces. Touch-screen interfaces, with features like on-line documentation help screens, and maintenance manuals are regular features as opposed to a few years ago when an on/off button with a few LEDs was common.
About the Author
Kenneth Gray is vice president of sales and marketing at Aseco, a position he has held since 1990. Mr. Gray earned a B.S. degree in industrial engineering from Northeastern University. Aseco, 500 Donald Lynch Blvd., Marlboro, MA 01752, (508) 481-8896.
Copyright 1997 Nelson Publishing Inc.