Microwave Integrated Circuits (MIC) and Monolithic Microwave Integrated Circuits (MMIC)
Around 1969 LEOK started the development of its own thin-film facility. The main purpose was to have a facility for the design, production and testing of Microwave Integrated Circuits (MIC). At that time, the design process was based on computer-aided design (CAD) techniques. In 1971 a start was made with the development of the first MIC components.
Examples of components developed in the MIC facility were: a directional coupler, a 3 dB coupler, Tchebisheff filters, a spiral antenna, a circulator and a balanced mixer.
The application of MICs was in transmitters, receivers and antenna systems of radar, communication and Electronic Warfare systems.
MMIC technology (1985 – 1995)
Monolithic Microwave Integrated Circuits (MMIC) technology has become an essential part of modern radar systems. In the 80s, based upon the knowledge acquired of phased array radars (FUCAS, TOEKAN) the radar research group realised that for even more flexibility and especially high reliability, active phased array radar technology could be the answer. For the realisation of such a radar, the application of MMIC was a possibility. MMIC creates the possibility of manufacturing complete microwave circuits in a single integrated circuit. Before making a real start with MMIC work at TNO-FEL, discussions took place with the Ministry of Defence, Hollandse SignaalApparaten (HSA) –now Thompson- and the Philips Physics Laboratory in Eindhoven. This resulted in an apprenticeship of a TNO staff member at Philips. From the end of 1987 until the beginning of 1989, he was trained on the job at the Philips “Laboratoire d’Electronique et de Physique (LEP later known as PML, Philips Microwave Limeil, and even later known as OMMIC) in Limeil-Brévannes near Paris. When he returned to TNO, a small group for MMIC design was set up under his guidance.
In 1988, one single TNO MMIC was taped out to one mask set. It was TNO’s first MMIC, realised in a 0.7 μm GaAs MESFET technology from LEP, with a transistor count of less than 10 on a single die and implemented as a building block for a vector modulator. The use of Gallium Arsenide (GaAs) had a number of important advantages. E.g. the material has the property of realising low noise circuits. Besides GaAs-MMICs are lightweight, compact and have high reliability and reproducibility. The GaAs technology moved fast. In 1989, a 0.7 μm MESFET was an outstanding technology, in 1992 a 0.5 μm MESFET was ruling, and in 1996, a 0.25 μm pHEMT device was offering lower noise and higher power densities.
For the realisation of MMICs, it is required to have facilities such as workstations, design software and test equipment. The production of the MMIC has to be carried out by a foundry like LEP. A foundry can produce wafers with a diameter of 50 mm, which can contain a large number of MMICs. For example 350 circuits of ten different types.
In 2014, 25 years later, TNO produced roughly 30 original MMIC designs on ten different mask sets in III-V and IV-IV semiconductor technologies. Not a single of these technologies even existed in 1988. If we include the commercial mask sets, the volume is too large and diverse to count reliably. Implemented circuits encompass complete multi-band receivers, high-power amplifiers and T/R core chips. Technologies include GaAs, GaN and SiGe. MEMS and ferrite integration are pursued in parallel. Challenges to overcome include the combination of electrical, electromagnetic, thermal and mechanical problems in as much a multiphysics approach as we can manage. Examples are integrated limiters with a limiting power of up to one kilowatt and the many-thousand transistor 8-channel consumer phased-array MMIC with integrated control.
In both the military arena and in other domains, the TNO radar group is renowned and in a number of radar areas, international experts even consider TNO to be the world’s leading radar group
In 2014, TNO houses a group of approximately 15 people, loosely denoted the MMIC group, with well over 250 man-years of experience in MMIC design for phased arrays. Six of these have worked on the subject at TNO for over 20 years. In these 25 years, TNO published more than 100 scientific papers on MMICs.
The impressive performance of the Dutch air defence and command frigates and Holland-class patrol vessels is, to a considerable extent, based on the MMICs that have been developed and tested by TNO
The MMIC research concerns three main themes:
- Research on individual microwave functions that enable active phased-arrays in the first place.
- Research on highly integrated core chips, in combination with high-power, high-efficiency power amplifiers that (1) simplified T/R module design dramatically and (2) enabled more than 10 W at X-band for a single transmit-receive module.
- Research on integrated receivers for digital beamforming systems, and the corresponding technology re-partitioning.
To enable these main themes, real breakthroughs were made to model accurately passive and active components, to efficiently design complex linear and non-linear circuits at microwave frequencies, to characterise non-linear components under high-power conditions, to simulate non-linear and harmonic behaviour and to test efficiently large numbers of circuits.
MMIC design software
CAD software to design MMICs underwent major changes over time between 1988 and 2014. TNO acted as an industry leader in accommodating these: from home-written linear S-parameter-based microwave simulators, via Philips’ internal spice-like simulator PHILPAC, via the legendary Touchstone (netlist-based manipulation of S-parameters with graphical layout) via EEsof’s Libra introducing schematic editors and harmonic balance simulations to provide harmonic simulations), evolving into EEsof’s Series IV which included methods to simulate non-linear behaviour in time- and frequency domain of transistors at microwave frequencies. These simulators would later merge with Hewlett Packard’s Microwave Design System into ADS and were complemented with electromagnetic solvers in the nineties (Sonnet, Momentum, HFSS and later many others contributed largely to the maturing of MMIC design).
Measurement systems were also not available. The chuck of the first probe station was moved manually, unimaginable from the current automated-probing-of-full-wafers perspective. The first load-pull measurement setup at TNO employed not only in-house software but also a vector modulator built around an MMIC that was developed in-house. We needed to manufacture an MMIC to advance the MMIC design state-of-art! The in-house developed load-pull system allowed impedances up to the outer border of the Smith Chart, thus enabling the design of a high-power and highly efficient power amplifier.
PHOTO The demonstration of 5-watt hybrid high power amplifiers (HPA) on a single die, illustrated in a ceramic package in the photo below, revolutionary changed the necessary aperture.
Between 1997 and 2003, optical techniques for phased-array application have been thoroughly investigated. The research questions included the possibility of antenna remoting over coherent or non-coherent links, and the possibility of optical true-time delay beamforming that had the promise of extremely wideband performance and optoelectronic RF generation, with the demonstration of an optical PLL as the research vehicle.