Millimeter-wave Radio Front Ends: The Present and the Future

Millimeter-wave Radio Front Ends: The Present and the Future

Lamberto Raffaelli
ARCOM Inc.
Salem, NH

Millimeter-wave (MMW) point-to-point digital radios have been in production for several years and, due to the high demand of rapidly deployable cellular telephone networks, their market is growing worldwide. This rapidly growing MMW wireless market is challenging the microwave industry with requirements for sophisticated MMW front ends manufactured in high volumes but with prices significantly lower than similar products produced just a few years ago. This article attempts to analyze the available market and required technology both today and in the near future.

The MMW Wireless Market

The Point-to-point Radio Market

The point-to-point radio market has grown rapidly from a niche business to a multibillion- dollar opportunity. These radios are mainly used in the cellular telephone network as a rapidly deployable link between base stations. The alternative is the traditional wired solution, which, if the fiber is not already down, is certainly more expensive and requires additional time to install. A point-to-point radio market summary is listed in Table 1 ; a conservative price erosion of approximately 10 percent per year is assumed. As described, the radio business already exceeds $2 B and will continue to grow at a rate of approximately 25 percent per year.

The Total Market

Point-to-multipoint implementation presently is in the field trial stage and should move to production within the next two to three years. Even though this market has the potential to grow exponentially, performance and cost issues still exist that could delay its real market penetration. In any case, the point-to-point business is expected to remain the largest MMW opportunity for the next five years. Other MMW market opportunities exist the area of satellite communications and sensors, including devices such as automotive radars and speed sensors. Details of the overall MMW market are listed in Table 2 .

Transceiver Design

The current technology for today's digital radios is focused on two distinctly different areas. The first area involves low cost, high volume, less sophisticated frequency-shift-keying (FSK) radios primarily used in PCN. These radios are somewhat stripped down from the previous FSK generation. Figure 1 shows the typical block diagram for a low cost FSK radio. This design includes a single synthesizer, low output power and moderate noise figure. The synthesizer output frequency is primarily in the 2 to 13 GHz range. This lower frequency input allows for the use of inexpensive surface-mount technology (SMT)-type components to amplify and multiply the LO signal that feeds the transmit (TX) upconverter and receive (RX) downconverter. The modulated signal is upconverted and then amplified to the desired output power (typically < +15 dBm). Power control is typically achieved through bias control on the output amplifier. The receive function is handled by a single-balanced mixer.

Higher frequency (> 26 GHz) radios do not contain a low noise amplifier (LNA) at the input in order to reduce cost. However, the missing LNA leads to noise figures in the 7 to 10 dB range. Some lower frequency radios are able to take advantage of low cost SMT LNAs and offer noise figures of < 5 dB. Much of the assembly of these low cost radios is done on surface-mount pick-and-place machines, which helps to reduce the overall manufacturing costs significantly.

The second focus area is currently in lower volume, sophisticated quadrature amplitude modulation (QAM) radios. These radios are more expensive than the earlier FSK radios, but offer higher data rates and better band efficiency in some of the already crowded frequency bands. The one similarity the radios share with the low cost versions is the single synthesizer used for the common LO. The synthesizer is where the most effort has been expended in developing these QAM radios. Its performance, along with the ability to generate higher output power and state-of-the-art noise figure, defines the radio's architecture.

Figure 2 shows a block diagram of a typical QAM radio. At first glance, the radio appears to be quite similar to the FSK radio, but some subtle differences exist. First, on the TX side, there is typically a double-balanced upconverter. This configuration is required to improve linearity and minimize LO leakage at the RF port. Also on the TX side, a power amplifier is added at the output. The output power stage is used to obtain P1dB power levels of  greater than +27 dBm and IP3 performance of +36 to +38 dBm. Power control is accomplished through attenuators at the upconverted frequency to avoid TX noise figure problems. On the RX side, an LNA is added before the mixer. Many systems require state-of-the-art noise figure performance from the LNA. Care must be taken to balance the overall RX noise figure and the input P1dB. The remaining components are similar to those used in the low cost FSK-type radios.

Manufacturing Technology

Two key requirements in this dynamic business are low cost and speed of execution across the entire design and manufacturing cycles. Therefore, design time and production ramp-up periods must be shortened to succeed in this market. In addition, customers are constantly looking for companies capable of reacting rapidly to changes in engineering and production schedules. As is well known, manufacturing costs comprise several components: materials, labor and overhead charges. In order to reduce costs, it is imperative to reduce not only materials and labor, but also the overhead and administrative infrastructure. Since MMW technology was originally developed by military contractors, most of the companies trying to win commercial communications contracts must work hard to optimize organizations that are too expensive to compete in the commercial MMW communications market.          

It is not only a problem of cost, but also an issue of speed. A management-heavy organization designed to deal with long-term steady military contracts is usually unable to move quickly. This handicap could be a major impediment in the commercial market where the decision process needs to be extremely fast. From a cost point of view, volume is the key driver. Since materials represent 40 to 50 percent of the overall price, large volumes allow the introduction of low cost packaging technology. The package, including connectors and DC feeds, represents approximately 30 percent of the material bill. For a MMW package, die-casting has a break-even quantity of several thousand units. Injection molding could be even more cost-effective, but requires larger volumes. Overall, the ability to utilize common parts through all architectures is the single most important factor in reducing cost.

This component commonality also reduces the risks associated with inventory. From a labor point of view, automated assembly processes are mandatory as soon as the volumes reach levels of several thousands. An automated assembly process creates instant capacity on the manufacturing floor and helps reduce production ramp-up times.

All this capacity is wasted if the bottleneck moves to the test area. At MMW frequencies, the use of tuning to compensate for process and semiconductor variations is still very common. Since tuning automation is very unlikely, the only way to reduce cost is to remove or significantly reduce tuning time. This reduction in tuning time is achievable only if the design allows for manufacturing tolerances. The completion of some production runs is usually required to properly understand where the limits are and to adjust the overall design to make it a high yield MMW product. If the volumes are as large as predicted for point-to-multipoint systems, custom MMICs could also be significant in reducing cost. Figure 3 shows a 26 GHz QAM transceiver that utilizes an automated assembly process.

Conclusion

The MMW commercial market is offering unprecedented opportunities to the microwave industry. Communications and other applications will double this multibillion-dollar business in the next few years. To be successful, companies must focus mainly on flexible manufacturing and overall speed of execution from design conception to product shipment.

Reference
1.         L. Raffaelli, "MMW Digital Radio Front Ends: Market, Application and Technology," Microwave Journal, Vol. 40, No. 10, October 1997, pp. 92-96.

Lamberto Raffaelli received his laurea degree in electronics engineering from the University of Bologna, Italy in 1975 where he remained as a researcher of CAD systems for instrument landing systems until 1977. In 1977, Raffaelli joined the Microwave Division of Elettronica, Rome, Italy, as solid-state chief engineer, designing GaAs FET amplifiers and other MIC components and subassemblies. In 1983, he joined Teledyne MEC in Palo Alto, CA as a staff engineer, developing 2 to 18 GHz GaAs FET amplifiers. From 1984 to 1994, he worked for Alpha Industries in Woburn, MA where he served as director of MMW commercial products in the automotive and communications areas. In 1994, Raffaelli founded ARCOM Inc. to design and manufacture MMW wireless products for digital radios and other applications.