MMW Digital Radio Front Ends: Market, Application and Technology

Lamberto Raffaelli ARCOM Inc. Salem, NH

Point-to-point mm-wave (MMW) digital radios offer unprecedented opportunities to the microwave industry. This technology, which was developed mainly for military applications, is finally finding a good market in the commercial arena. Along with automotive radars, wireless television applications and other MMW opportunities, the digital radio market soon may be able to transform a company with only a few expensive prototypes into a multibillion-dollar business comparable to the low frequency wireless market. This article describes the application, market and technology required for MMW digital radio front ends.

Market Opportunities

The opportunity for growth at MMW frequencies is unprecedented. The digital radio market originated in Europe and is now growing rapidly in East Asia and the US. At first, most of the volume was developed in the 23 and 38 GHz bands. Industry leaders such as Alcatel, California Microwave, Digital Microwave, Ericsson, P-Com and Siemens are producing several thousand radios per year and are projecting an annual growth of more than 25 percent. Other frequency bands that are promising to grow rapidly and take on a large share of business are the 12, 15, 18 and 26 GHz bands. This radio market is believed to be close to $1 B already. According to a projection by Mark Thorn of Hewlett-Packard Technology Center, Santa Rosa, CA, there is currently a market of 90,000 radios for the 23, 26 and 38 GHz frequency bands.

Table I

The Radio Market







Price erosion






Number of radios






Trans. price($)






Market for transceivers ($M)






ODU price* ($)






Market for ODUs ($M)






Radio price** ($)






Market for radios+ ($B)






* ODU includes the transceiver, diplexer, synthesizer(s), and IF processor

** Radio includes an ODU and IDU (mainly the modem)

+ Forcast assumes a conservative price erosion

The market forecast listed in Table 1 assumes that these three frequency bands comprise 65 percent of the business and a current transceiver, outdoor unit (ODU) and overall radio price of $1000, $2000 and $6500, respectively. By 2001, the market will exceed 200,000 radios, as listed in Table 2 .

Table II

Radio Market Forcast (Thousands)


23/26 GHz

38 GHz

















Microwave radios have long been used for transmitting data and voice with carrier frequencies in the 2 to 15 GHz range and data rates up to 274 Mbps. A complete radio comprises an RF front end and a processing section. The RF front end represents more than 50 percent of the overall radio cost. Long-haul (up to 25 miles), high data rate, point-to-point radios are used for deploying long-distance communications networks. Due to the output power requirements, high level of complexity and small volume, each radio's associated cost is high (as much as $200 K). Traditionally, these radios have used analog modulation techniques and, therefore, are classified as analog radios. More recently, digital radios (referring to digital modulation techniques) have been developed and deployed at 18, 23, 26 and 38 GHz carrier frequencies to provide short-haul communications (10 miles or less) for private networks and for the interconnection of mobile telephone cell sites. Digital radios offer improved transmission quality compared to analog radios due to their inherent noise immunity. In order to improve cellular telephone service, the number of stations in the personal communications network (PCN) must increase dramatically and the information technology must advance from analog to digital. The connection between stations could be realized with a traditional wired solution such as wire cable or optical fiber. However, the cost of deploying a cable in a densely populated area ($40 K/mile) gives wireless links, which sell for approximately $14 K, a significant economical advantage. In countries where the telephone network is obsolete, the PCN MMW radio infrastructure offers a low cost and rapidly deployed alternative to the traditional cabled network. In addition to point-to-point radio links, the point-to-multipoint application for television channel distribution and wireless local areas also is moving rapidly from research and development to production. This market has the potential to grow at least 10 times as large as the digital radio market.

Required Technology

First Generation

The first-generation MMW front-end, shown in Figure 1 , is based on waveguide components and Gunn-based fundamental oscillators. The Gunn device was directly frequency-shift keying (FSK) modulated. A manually set waveguide attenuator was used to adjust the transmitter output power to the actual link requirements. At 38 GHz, the typical transmitted output power and receiver noise figure at the antenna port were +15 dBm and 12 dB, respectively. On one hand, the Gunn-based oscillator offers optimum phase noise and the waveguide-balanced downconverter presents good isolation. On the other hand, with minimum integration, the size and cost certainly are not optimized. In addition, to cover the entire bandwidth, several subband options are required, which complicates the inventory management. Gunn oscillators also have proved to exhibit microphonic and reliability problems.

Current Generation

In order to reduce the subband options, the current-generation source is based on a bipolar transistor oscillator multiplied up to the desired frequency and bandwidth. At 38 GHz, multiplication factors range from 6 to 20. Due to this multiplication, the entire bandwidth now is covered with one option. By using planar transmission line fabrication techniques, the integration level is optimized, reducing the front end's size and cost. The current-generation front end, shown in Figure 2 , also takes advantage of 0.25 mm psuedormorphic high electron mobility transistor-based GaAs monolithic amplifier technology, which, due to the reduction of bond wires and level of assembly complexity, results in significantly improved yields and reliability. Another advantage to this solution is that the transmitter output power adjustment is realized electronically and, therefore, can be set by software. This function simplifies the installation process significantly because operators are no longer required to perform awkward power measurements on the tower. However, the large VCO bandwidth causes a phase noise degradation and the multiplication process generates harmonics and products that are sometimes difficult to remove. The current-generation manufacturing process uses little automation. This market is certainly growing, but not without painful and frequent changes, and the volumes are not as large as they appear. Because the quantities per unique design do not exceed a few thousand per year, a highly automated line remains inefficient. In the future, due to the much larger projected volumes, increased levels of assembly and test automation will become more cost effective.

Next Generation

In order to improve cellular telephone service, the PCN network is expected to grow, utilizing a large number of base stations and, therefore, a larger number of shorter links. As a result, a high volume of low cost, less sophisticated FSK radios will be required. However, the market also will require higher bit rates and more band-efficient quadrature amplitude modulation (QAM) radios. Due to the high linearity requirements, higher output power capability and more critical synthesizer, these radios are more complex and more expensive. Table 3 lists a comparison of the performance required at +80°C from the low noise and power amplifiers for FSK and QAM radios.

Table III

Amplifier Requirements at 80C

Frequency (GHz)

Noise Figure (dB) FSK and QAM

Psat (dB) FSK

P1dB (dBm) QAM


< 3.0

20 to 25

25 to 27


< 3.0

20 to 25

23 to 27


< 3.5

20 to 24

20 to 25


< 4.0


20 to 25


In the point-to-point MMW radio market and in other MMW commercial markets, including automotive radar and wireless television, the expected production volumes are likely to transform a business consisting of a few costly prototypes to an industry of commodity products. Planar technology has made waveguide Gunn-based front ends obsolete. Monolithic technology, which was developed mainly for sophisticated military systems, is finding real opportunities and a good market. Because the most important parameter for success is cost, the MMW industry must transform itself from a prototype shop to a mass-production consumer business. This industry transformation, which has been expected for years, is finally becoming a reality.

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.