WARRINGTON, Pa. — Citing the lack of EDA tools for RF design, the International Wireless Packaging Consortium (IWPC) is hoping to spearhead the development of a tool set for cell-phone RF package design. The consortium includes representatives of 36 companies in six countries, including Nokia Research Labs. The group's last meeting, in fact, was held in Helsinki.

Rather than cell-phone ICs, the IWPC initiative focuses on the packaging and interconnection materials used in the phones' radio-frequency front ends. Though handset manufacturers like Nokia and Qualcomm are involved in the effort, along with EDA vendors like Ansoft, HP EEsof and Sonnet Software, the bulk of IWPC support comes from material, packaging and interconnect suppliers like BF Goodrich, Bourns, DuPont Electronic Materials, Heraeus Cermalloy, Kyocera America, Merrimac Industries, Rogers Corp. and Sheldahl.

IWPC hopes to advance the work begun with the Darpa-financed Microwave and Analog Front-End Technology program. That $59 million project hoped to standardize the tool sets and interchange formats for RF semiconductors, and was applied to high-GHz antennas and satellite links. The IWPC initiative, in contrast, will concentrate on substrates and interconnect materials, and will attempt to aid cost-sensitive applications like cellular handsets and wireless LANs.

Already, as another project, the IWPC is sponsoring workshops on low-cost phased-array antenna technology that will support consumer television applications.

The IWPC believes its Electronic Product Design System (EPDeS) will be the only layout scheme to focus on the geometry and materials used to package RF cell-phone front ends. The goal is to provide designers with desktop tools and data for advanced trade-off analysis, said Don Brown, IWPC's director. A byproduct of the design system, of course, will be reduced unit costs and time-to-market. "The cellular phone makers are continually asking, 'where can we save $5?'" said Brown.

"The general idea of this design system is to integrate existing mechanical, thermal and electrical design tools across an open interface so that the data in one tool can be used by another tool in the system," Brown explained. Right now, few design tools can transfer the material- and geometry-sensitive layout information.

"The RF guys have screamed that the design tool guys don't get it, that they don't understand the problem," said Brown. "That situation has obviously improved, but there still is a gap between the needs of the RF designers and what the design-tool makers can deliver."

At cell-phone transmission frequencies-800, 900 and 1,800 MHz-the electrical traces and their spacing from each other must reflect some multiple of the wavelength of the transmission frequency. Otherwise, severe attenuation and other debilitating effects (like crosstalk and interference) will wreck the RF signal. This is increasingly important as the battery-powered handset, in urban areas, must send and receive relatively puny signals in the presence of more powerful RF stimuli.

Interconnect and substrate materials will have similarly profound effects on signal quality. Some manufacturers are moving from co-fired ceramics to multi-layer phenolic materials as a means reducing the size, cost and weight of RF modules, but success in this area-especially for consumer electronics-will depend on tight control of packaging geometries over time, temperature and frequency.

In fact, a major obstacle of the EPDeS, says Brown, is an interchange format for geometry information. Much of the information, ordinarily used to describe thick-film interconnects on ceramic substrates is formulated in "mils" (milli-inches)-that is, on a layout grid that reflects a tolerance of 10-3 inches. In new-generation layout tools, that tolerance needs to be 10-6 inches, said Brown.

An additional problem is the simulation and placement of passive components-capacitors and inductors-used for tuning, re-tuning and fine-tuning in the cell-phone's RF module. Kari-Pekka Estola, vice president of Nokia and director of its electronics laboratory, told engineers at a the Wireless Communications Conference in San Diego last month that, while the number of ICs was shrinking, the typical cellular handset of the year 2000 will include as many as 100 passive components .

These passives, combined with antenna and shielding requirements, said Brown, contribute to the packaging costs of a cellular handset. By offering designers access to detailed models of materials and manufacturing processes, the IWPC hopes to create a smoother pathway from schematic to packaged module.

Like IEEE subcommittee efforts, the IWPC is depending on the voluntary efforts on its members to define and develop the EPDeS. Still, Brown sees the cell-phone manufacturers in his group as major drivers of the effort, and believes that a usable system will come together within two years. The kickoff meeting of the EPDeS working group was held in Helsinki in October. The next meeting will be held Feb. 26-27 in San Diego, Qualcomm's hometown.