There is a current trend in microwave and millimeter-wave system development toward higher-levels of integration on a single chip, motivated by a desire to reduce cost, size, and complexity, particularly as many old applications are being reinvented in array architectures. Thus, high-performance Monolithic Millimeter-Wave Integrated Circuits (MMICs) of all kinds up to about 50 GHz are readily available at low-cost from commercial suppliers. The selection of commercial MMICs above 50 GHz is more limited, but new ones are introduced regularly, and the availability of foundry services for III-V semiconductors from several companies makes it possible for us to develop custom MMIC designs for our own unique applications well over 100 GHz.

At the CDL, we are developing custom MMICs and MMIC-based Multi-Function Modules (MFMs) for a vast array of projects, like ALMA, EVLA, DSN Array, SKA, and FASR. Some recent examples are highlighted below:

1. Schottky Diode MMIC Multipliers and Mixers for ALMA
   

   The LO subsystem of the ALMA telescope must generate significant power in phase-locked loops at W-Band to drive cooled submm-wave multipliers in the front-end. This has required the development of a complete set of MMIC multipliers and mixers ranging from about 65 to 145 GHz [1]. Each of these GaAs Schottky-Diode components has been fabricated by United Monolithic Semiconductors. For example, below are the two custom non-linear MMICs designed for Band 6, which operate from 74.3 - 87.7 GHz.

Fig. 1. MMIC tripler for ALMA Band 6. Output power is roughly 1 mW from 74.3 - 87.7 GHz. Chip dimensions are 2.0 x 0.73 x 0.1 mm.

Fig. 2. Double-balanced mixer for ALMA Band 6. Conversion loss is 10 - 15 dB with LO power of +19 dBm. Chip dimensions are 2.0 x 0.73 x 0.1 mm.




2. Power Amplifiers for ALMA
   

   The ALMA W-Band multipliers described above generate roughly 1 mW of power in the designed frequency range. However, much more power than that is needed to drive the cooled submm-wave multipliers that follow in order to get a strong enough LO signal to the SIS mixers in the front-end. Therefore, a set of MMIC Power Amplifiers is also under development for this project. Implemented both in InP and GaAs, these PAs are required to deliver above 50 or in some cases 100 mW in their respective frequency bands. This represents state-of-the-art performance for millimeter-wave PAs with this frequency and bandwidth [2].

3. Compact Water Vapor Radiometer for the EVLA
   

   This is a MMIC-based module under development for the EVLA. It was designed as an attachment to the existing K-Band receiver to take advantage of the high-sensitivity front-end. The entire 17-26.5 GHz band is amplified, filtered and split into five frequency channels which are then detected in order to measure the 22.25 GHz water vapor line in the atmosphere, and provide phase corrections for the rest of the array. A block diagram and photograph of the module appears below. The module includes switches for selecting the antenna polarization and removal of "dark-current" in the detectors, and digitally controlled step-attenuators for leveling the power across all 5 channels to account for gain slope and ripple that may be unique to each front-end.

   Due to the relatively low-frequency of this component, it was possible to use all commercial MMIC chips. The total replication cost of the module is just over $500. In addition, the smaller size afforded by adopting a MMIC approach for this module will make it easier to meet the tight temperature-stability requirements. The smaller intrinsic size is also what makes it practical to include 5 channels instead of 3 for greater fidelity in the water-vapor measurement.

Fig. 3. Block diagram of the prototype CWVR MMIC-module.

Fig. 4. MMIC Multi-Function Module for the CWVR

4. Wideband MMIC components for the SKA
   

   A key advantage of the US Large-N/Small-D proposal for the Square Kilometer Array of other designs is the ability to reach higher frequencies. The combined requirements of extending this frequency coverage into the low millimeter-wave band and minimizing the global cost of the array leads to two conclusions. First, we must reduce the number of independent receivers needed by developing very wideband receivers, and second we must optimize the tradeoff between cost and performance by selecting a MMIC approach for as many elements of the receiver as possible.

   The highest frequency band in the US SKA proposal is 11 - 34 GHz. A MMIC LNA has been designed for this band and is currently being tested and revised [3]. The initial results on this chip are shown below.