|
Testing Integrated RF Devices: Keeping Ahead of the Technology Curve
|
The current technical challenges are no different from that of the 1980s and 1990s when there were many new test vendors claiming that they had the ability to build a low-cost, mixed-signal tester that met all the requirements of their higher-cost competitors.
The "big boys" who introduced higher CPU speed testers with higher accuracy, as demanded by OEM end-customers, eliminated most of these vendors.
Low-cost test equipment vendors were not able to capitalize on an economy of scale in manufacturing high-speed digital testers with high pin counts. There was also difficulty in synchronizing the digitized analog instruments with the digital test patterns used in testing mixed-signal devices. Low-cost test vendors, however, are now able to accomplish their goals with today's semiconductor technology.
Analog Options
On the other end of the equipment spectrum, established high-end, mixed-signal test equipment vendors are adding RF instruments or RF add-on analog options to their testers, and are staying competitive based on multisite testing, test accuracy and test repeatability as demanded by the emerging integrated RF devices.
The Challenge of Market Requirements
With continued price erosion of emerging integrated RF devices into commodity components, most OEM end-customers are favoring the use of low-cost RF testers.
One current trend is toward sufficient characterization of new integrated RF devices using high-end mixed-signal testers. Other trends include understanding design-wafer process marginality and its sensitivity to testing. Also important is the determination of sufficient test coverage using low-cost RF testers for mass production.
Past history reveals that products that "erode" into commodity devices usually leave behind a trail of casualties.
In this case, the number of customers who were able to weather the price battle contracted into a handful of independent device manufacturers (IDMs). They survived due to the economy of scale of mass production volumes with reasonable wafer, package and test costs.
Subcontractor houses were engaged to meet these low-cost, mass production volumes. Some contracts were even based on the mathematics that overall costs must be lower than the cost of the IDM's in-house manufacturing.
Balancing Act
It is always a struggle to balance the production of products that are in emerging and growth stages (moderate volume with moderate testing cost) with products in the mature and declining stages (high volume and low testing cost) of their product life cycle. The challenge is to identify who will be the winners in the pool of customers launching and growing their new products.
One tool often used for decision making is the gross profit margin (GPM) of the individual product testing. This may not be a reliable measurement, however, as illustrated in Figure 5.
As the figure indicates, to maintain similar total revenues, a subcontractor must ensure that there is at least a 10X ~25X increase in volume for mass production.
With the additional number of low-cost test systems needed to generate the equivalent test revenue, the total investment may be much higher. These factors will definitely affect the overall factory floor size, utilities and the manpower needed to support test operations.
The scenario of the "big guys beating out the smaller ones," and the illusion that low-cost test systems are cheaper, is often underplayed in the decision-making process by other, more compelling real life factors.
In many situations, a wide range of considerations affects initial decision making, which is often colored by emotional reactions to a vendor or a vendor's reputation. Extraneous factors that can affect decision-making may include the following scenarios:
Vendor failed to perform well on an earlier RF project: "I hate Vendor A because it was like hell trying to develop RF test program on that platform."
The proven RF experience and exper-tise of one vendor over another: "My previous company only buys Vendor B's testers. It has compelling engineering reasons and experiences to confirm that Vendor A's testers are not suitable for RF."
Hearsay evidence of a vendor's failure to meet another company's RF specifications: "Do you know that many companies are returning testers to Vendor A for not meeting specs?"
Vendor must be customer approved: "The customer wants us to buy Vendor B's tester . . . otherwise, we'll lose the business. It will also give us a foothold into this strategic account."
Vendor provides excellent support and additional "benefits": "Vendor B has very good application support . . . and offers many freebies!"
Regardless of the criteria used, making the final decision is often an uphill battle. It is ultimately based on calculated risks and justifications in providing a cost-effective test solution without compromising test coverage integrity and test quality.
| |
Discrete Devices |
Integrated Devices PC & Comms |
| Customer ASP$ |
Below US $1 |
US $3-$20 |
| Test ASP per device |
1¢-4¢ per unit (1X) |
25¢-40¢ per unit (10X) |
Volume loading to achieve equivalent gross revenue |
Volume = 10X-25X |
Volume = 1X |
Capital investment (per ATE tester) |
US $300K (1X) |
US $1M-$2M (3X-7X) |
Total investment for equivalent revenue* |
US $3M-$7.5M |
US $1M-$2M |
| * Assume product offload (no test development cost); equivalent GPM and test-time per device; and similar
|
Conclusion
Mounting consumer demand for faster and better products means RF manufacturers are increasingly sensitive to time-to-market and cost-of-test pressures. Further complicating matters, these manufacturers must move their next-generation products to market on the first try or risk being left in the dust by the competition.
The role of test becomes eminently important in this environment, requiring savvy manufacturers to use the most reliable and cost-efficient RF testing solutions available.
* Bluetooth is a trademark.
 |
| Nelson Lee |
Mr. Lee is STATS' Director of Test Technical Support. He was instrumental in establishing the RF test capabilities to strengthen STATS' core competence in testing next generation communication devices.
With over 20 years' experience in semiconductor test, Mr. Lee was previously with TriTech Microelectronics Ltd. as Director of Test Development. He was also formerly at Chartered Semiconductor, Sierra Semi-conductor, Telmos Inc. and Fairchild Semiconductor.
[leenelson@stats.st.com.sg]
|
|