Issue



How Advances in RF and Radio SiP Affect Test Strategies


08/01/2005







Integration requires a high level of test coverage across SiPs

BY MARK BERRY

The final RF SiP module test can require specific, protocol-related system tests, in addition to conventional ATE tests. These system tests usually have been performed by OEMs or their EMS providers. The outsourced semiconductor assembly and test (OSAT) industry is well-positioned to support the combination of advanced assembly technology and the need for a complete test strategy.

In the last few years, significant advances have been made in SiP technologies and processes (Figure 1). These advances leverage the strengths of advanced IC assembly/packaging technologies, and have moved beyond typical EMS, such as mounting packaged ICs and components to a PCB. Leading OEMs have begun outsourcing final assembly and system test. Examples include cellular phones, where leading cell phone manufacturers have outsourced to EMS the actual manufacture, in-circuit-test, and final system test of the handset, allowing OEMs to focus on more value-added core competencies. With increased use of SiP modules, more of the final product content is set to come from OSAT companies. Examples include integration of bare die, flip chip, and package-on-package in current SiP products. Other examples, even more specific to RF applications, include embedded RF shields, SAW filters, and support for 01005 surface mount devices. These new SiP technologies are based increasingly on chip-and-wire and flip chip technologies, which have been the strength of OSAT assembly houses. Often with such high integration there is only one RF port for transmit
eceive, one digital I/O, control pins, and power supplies. Within the next 2 years, embedded antennas will also become commonplace. These highly integrated SiPs assist an OEM in radio-module re-usability across many handset models and provide higher yields in the handset factory.


Figure 1. A SiP is a fully integrated system or subsystems functional block that bridges the gap between SOC and SOB (system on board).
Click here to enlarge image

While SiP radios and front-end modules (FEMs) can bring greater simplification and cost reduction to the final product manufacturer, there are many challenges in test. On one hand, the combination of various functions and technologies in a SiP can bring about more simplified ATE requirements. However, with such increased levels of integration, the test coverage required is heading more toward testing the end application.

In advanced cellular SiP applications, a front-end transmit module can be integrated into a SiP with the transceiver, baluns, and filters to make a complete radio front-end that readily connects to the baseband processors (call modem). Traditional RF/IF ATE tests are well-established for the transceiver and FEM power amplifier. Usually, the OEM or EMS provider receives these tested, good transceivers and FEMs, along with the rest of the bill of materials (BOM), and then proceeds to assemble the phone PCB and perform system tests. With an integrated SiP, it is more sensible to move these system tests into the OSAT. The system tester, typically a base-station emulator, can be used as a quality or “gate” check in the OSAT’s high-volume ATE test line to complement the testing. Tests such as time mask, phase error, error-vector magnitude, fading, and bit-error rate can be too time-consuming to implement on an RF ATE tester. The system tester can also run proprietary code that the OEM supplies.

Advanced connectivity radio modules (802.11, 802.15, UltraWideBand), which offer complete integration of the baseband processor and RF front-end, are pushing the envelope of traditional RF ATE test. In these cases, ATE test is not supplemented with traditional cellular-type system testers, but rather the RF ATE equipment has to be mated with external measurement equipment and techniques; for example, up/down conversions must be relied on heavily. Examples of device performance which need to be tested include:

  • wide bandwidth (up to 22 MHz, future standards may require 40 MHz);
  • timing and amplitude accuracy (<1% amplitude error, <0.5° phase error);
  • support for high-speed interfaces, such as PCI and PCI Express.

When a reference crystal is integrated, its accuracy must be measured in ppm and accounted for in other tests. In industry standard 802.11n, measurement of signals up to 10 GHz in production test (20 GHz in engineering) are expected, but few RF ATEs support direct measurements above 6 GHz. In UltraWideBand (UWB) products featuring orthogonal frequency division multiplexing (OFDM), RF ATE testers and current bench equipment lack the IF bandwidth of the signals that need to be captured. Some of the UWB parts must capture 528-MHz channel bandwidth. A swept-frequency continuous-wave technique can be used, but it is time-consuming. In these connectivity protocols, there is less demand for sampling the production test with a system tester, and more demand for supplementing the RF ATE with extra hardware on the loadboard and/or extra instruments connected to the tester.

Therefore, advanced cellular and connectivity SiPs require complex final-module test strategies that can include EMS-style system testers and/or additional external instrumentation on the RF ATE. OSAT companies are well-suited to support these demands and work upstream of the final-module test to help customers achieve their cost, cycle time, and quality goals. Let’s review the upstream activities that occur in new product development and formulating the overall manufacturing and test strategy.

The furthest upstream activity is design-for-test (DFT). In SoC, this can be more straightforward than SiP. While individual ICs within the SiP have their own DFT features, the overall SiP module is usually a fast time-to-market, pieced-together mix of these ICs. Wherever possible, more emphasis must be placed on SiP DFT specifications upfront. Other product design considerations for test include thermal issues and pin mix. Thermal issues should be modeled and well-understood for both applications use and special test modes. Pin mix (digital, RF, analog pins) is an important consideration in selecting affordable ATE test equipment.

Next, consider the overall test strategy and manufacturing plan. This includes wafer probe, component procurement, component test, and final module SiP test. OSATs are usually positioned to provide this overall “turnkey” analysis and work with customers to develop effective test strategies. Some of the trade-offs to analyze include:

  • the amount of test coverage that can be obtained in the wafer-probe environment;
  • whether or not to package and test components within the SiP;
  • whether or not accelerated life testing should be performed, and to what degree (the whole SiP or just key components);
  • whether or not strip-test techniques are applicable;
  • which partners/suppliers should perform what tests, and so on.

In all cases, the costs of equipment (handlers, probers, testers, fixtures, interface hardware) and test times must be scrutinized to permit assessment of the overall test costs, all the way up to final-module SiP tests.

In wafer probe, OSAT companies are well-positioned to provide RF probe, when it makes sense for the end SiP, and to provide advanced probe techniques for bumped die and die that will be “flip-chipped.” For instance, with greater power-amplifier integration, 50-Ω matching is possible at the power-amplifier die inputs and outputs, as matching networks become integrated on-chip. This allows meaningful RF probe, which can catch parametric yield losses and save significant costs by pushing final SiP module-test yields up. Vertical probe and precise thermal management are two other important techniques for probing SiP components.

Yield can be maximized by using advanced probe and pre-tested packaged components in the SiP (“known good die”). When the probe and/or pre-tested package component test is performed by the same OSAT that tests the final SiP, then the OSAT can apply rigorous yield analysis and, where applicable, make proactive changes in probe or component test to improve SiP yield.

In SiP assembly, the OSAT can record full genealogy of the components used and create top-side marking codes that are useful starters and decoders for low-yield analysis of final-module SiP test. The OSAT can also take on procurement logistics across the BOM. Fast learning cycles are essential during assembly processes and SiP package development/qualifications. Choosing an OSAT with co-located assembly and test, and on-site failure-analysis capability, is advantageous to accelerate these cycles.

Conclusion

New advances in SiP technology favor the core assembly and package technologies of leading-edge OSATs. These new technologies make it possible to integrate virtually all transceivers, power amplifiers/FEMs, baluns, filters, etc. to create a complete cellular radio. In connectivity applications, the baseband can also be integrated, making a complete system SiP. With the increased functionality and integration of leading-edge RF SiPs , known good die and/or known good components must be used for the RF SiP to be cost-effective. OSATs that offer advanced wafer probe, component assembly-and-test (ICs), and RF test are well-positioned to deliver on these requirements. The OSAT can work closely with designers and product/test engineers to deploy the right test strategies across probe, component test, and module test to ensure results. The OSAT industry is well-positioned to integrate several traditional services performed by EMS providers and provide a complete turnkey service. This can reduce costs, speed time-to-market, and provide a faster feedback loop for continuous improvement.

References

For a complete list of references, please contact the author.

MARK BERRY, RF/test product manager and a senior director, may be contacted at Amkor Technology, Inc., 1900 S. Price Road, Chandler, AZ 85248; 480/821-2408, Ext. 5449; e-mail: [email protected].