Remember that exotic TV set your folks bought 30 or 40 years ago? They went to the dealer, looked at the choices, then spent a fair amount of time deciding whether the benefit was worth the high cost. Once the decision to buy was made, the dealer came out to install it and showed you how to use and adjust it. And as with many new technologies, the TV repairman became a familiar figure around the house.
Fast-forward to today. TV sets are inexpensive, easy to justify and available off the shelf; don’t require dealer installation or training, and live a long and happy life with little support or maintenance.
The world of ATE is following this pattern. Starting with the big, expensive ATE of the ‘70s and ‘80s, the trend is toward modern, compact, less-expensive yet powerful ATE. We are at—or nearing—the point where ATE is like a shrink-wrapped commodity: You choose what you want, take it to your plant and start using it like the TVs of today.
Trends in Manufacturing
Changes in manufacturing technology are both requiring and providing the opportunity for commodity ATE. Cell-based manufacturing provides workers with greater job satisfaction and allows the company to produce a variety of better-quality products. This manufacturing methodology places some special requirements on ATE:
A need for multiple ATE, one at each cell site. For this to be practical, the cost of multiple ATE must be affordable.
Flexibility. Users must be able to quickly change from one type of unit-under-test (UUT) to another.
The capability to share test data among systems for standardized, controlled testing plus the capability to support shared management production and statistical process control (SPC) reporting. This data sharing is best handled by standardized, low-cost network technology.
Not only are manufacturing trends forcing ATE to fit into these new molds. Other trends are also making low-cost commodity ATE practical.
Improvements in component quality reduce the need to test components for proper functionality at incoming inspection. High test coverage can be obtained by assuming the components are good, and by testing to ensure that the components are installed properly.
Automated manufacturing technology, such as pick-and-place machines, improves the process, but makes it more important to find faults as soon as possible. Also, the familiarity with PCs makes use of PC-based automated testers less intimidating. And, the ready availability of computer-aid-design (CAD) data for UUTs reduces lead times and the cost of customized test fixturing and programming.
Trends in ATE
Many trends in ATE technology lead to lower prices and allow a commodity approach to buying.
Test System Electronics
As with electronic assemblies, newer, efficient technologies provide a general price reduction in ATE. For example, look at the downward price trend in HP board ATE: starting at the high-priced HP 3065 to the HP 3070, continuing down in cost with the HP 3070 Series II.
Changes in Test Requirements
With the improvement in component quality, the trend is toward finding only process faults. This dramatically reduces the cost of a test system while sacrificing only a minor part of total fault coverage.
Surface-mount-technology (SMT) changes the fault spectrum from mostly shorted connections (through-hole technology) to open connections. New vectorless testing technologies can readily detect open-connection faults.
Changes in Test Technology
To find open connections prevalent in SMT (typically ICs and connectors), vectorless test processes use patented techniques with a stimulus applied to each pin to detect a parameter related to connectivity, such as capacitance.
Recently, some of these technologies have expanded to test the polarity of capacitors. These faults were previously detected with in-circuit test equipment. By using vectorless test, not only is the cost of the test system reduced, but the cost of custom fixtures and related programming and support is greatly reduced.
Use of the PC as a Test Platform
Using the IBM-compatible PC as a testing platform has reduced costs and proved efficient in several ways. First, the test electronics can be directly installed in a standard PC. This eliminates the need for external chassis and interface electronics, reducing system cost and increasing performance.
Second, standard PC networks easily share data for control of test programs and archival and analysis of test results. A plethora of software tools helps increase productivity. Programmer and operator familiarity with PCs shortens the learning curve for the test system.
High-volume production and a highly competitive market allow extremely low hardware costs for PCs. This lowers initial acquisition costs by an order of magnitude from earlier mini computers and workstations. And last, the prevalence of PCs makes repair, support and field upgrades of the computing platform quick and inexpensive.
Changes in Bed-of-Nails Fixturing and Programming
Advances in mechanical fixturing technology are also reducing costs and improving reliability. The traditional vacuum fixture is being replaced by mechanical and pneumatic fixturing. With this technology, fixture kits are less expensive, customization costs less (no tedious and problematic gasketing), and the fixtures last longer while in production.
Advances in software are making customization for each UUT less expensive. For example:
Standard CAD translators built into ATE reduce the programming time.
Tools such as Visual PCB (deMille Research) allow rapid processing of Gerber data to generate probe maps and diagnostic aids.
Interactive programming techniques reduce programming time. Early systems required lengthy compile cycles with each programming change. New interactive editors allow on-the-fly fine-tuning with SPC data readily available to set up program parameters and optimize performance.
Modern programming tools eliminate the need for specially trained programmers. Newer systems can use technician-level personnel with little or no formal training to perform fill-in-the-blanks programming.
Are All Applications Suitable for Commodity ATE?
Not all applications can be satisfied with commodity ATE. For example, in fully automated assembly lines, testing must be integrated with other processes. While commodity ATE can perform the testing, it typically must be tightly integrated with handling equipment.
Also, some types of testing are complicated enough to require specific hardware and expertise to do the job. For example, parametric test of ICs requires special high-performance ATE and ATE-specific expertise that is unlikely, in the short term, to be reduced to the commodity level.
An Example of a Near-Commodity ATE
As an example of a near-commodity ATE, CheckSum sells manufacturing defects analyzers and functional test systems. To demonstrate the commodity nature of testing, here’s a typical ordering/purchase sequence:
In almost all cases, sales are conducted over the phone, with supplementary technical information through the mail or the worldwide web. This provides direct engineer-to- engineer communications to reduce problems and lower costs.
Once an order is processed, shipment of standard product takes place in a day or two or up to four weeks when custom fixturing and programming are required. When received, the customer merely connects air or vacuum and standard AC power.
If service is needed, self-diagnostics can find the faulty module, and we overnight a replacement or loaner. Free factory training is offered, but only a very small percentage of customers request it. If new UUTs are to be accommodated, the customer requests us to build a fixture and program it, uses a third-party fixture vendor or undertakes the job in-house.
This simple commodity-type approach to ATE means initial costs are low, support costs are low, training and recurring costs are low, and implementation times are quick—all with little sacrifice of practical test coverage.
Summary
The ATE industry is taking the path of other new technologies. ATE’s history of high initial costs, special training, special expertise and high recurring costs is being replaced by newer near-commodity technologies.
About the Author
Brian Laine is the founder and sales and support manager at CheckSum. He previously held engineering and engineering management positions at John Fluke Manufacturing and Summation. Laine has a B.S. degree in computer science from Oregon State University and an M.B.A. degree from Seattle University. CheckSum, 19009 61st Ave., N.E., Building 4, P.O. Box 3279, Arlington, WA 98223, (360) 435-5510.
Sidebar
Benefits of Commodity ATE
Typical Sales Calls: None
Vendors to Deal With: One
Typical Factory Training Required: None
Low Cost: Typical System Costs $10k to $15k
Low Fixturing/Programming Costs: Typical Cost $1k to $3k per UUT Type
Quick Fixturing/Programming Times: Typical Lead Time of Four Weeks
Off-the-Shelf Systems: Typical Lead Time—In Stock to Three Weeks
Standard Network Compatibility
Statistical Process Control
Copyright 1997 Nelson Publishing Inc.
February 1997