Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
the design techniques are a little different. Form MMIC design we use more of S-Parameters design techniques for say designing an amplifier(ref gonzalez). However for the relatively low frequency RFIC designs we do use S-Parameters (but not for design fully). RFIC design such as LNA design is predominantly based on "low frequency analog circuit design techniques". We only simulate and find S11 and S22 and S21 as measures of i/p and o/p impedance and gain just because the test equipment can only measure s11 and s22 and s21 and not the direct i/p and o/p im pedances and voltage gains directly. I am pretty sure that we use S-Parameters only for sake of measurement. So i feel that MMIC being at relatively high frequency we begin to us microstrips and such "transmission line" effects start arising. If u check an LNA for 900MHz or even 5 or 6GHz using RF MOS u will not see microstrips or something...but rather only L,R,C...so design technique follows low frequency V, I circuit analysis technique. This is my personal view on this aspect. This is one aspect of RFIC design about which even I had some doubts always....We never seem to use the real S-Parameter design techniques for RFIC design for say an LNA....mainly because we can optimize in an IC.
The difference is RFIC initialy was monolitic IC that moving to microvawe frequency. MMIC initialy was hybrid microwave IC that moving to integration in one die. Now they are very close to each other and use the same design tools.
MMIC design is done on a different substrate, usually GaAs type materials.
The theory is almost the same. The tools are the same as well.e.g. ADS or Genesys suite etc. Sometimes you use more 'Momentum' type analyses while working on MMIC layouts.
This depends on the operating frequency, 1GHz = 30 cm(wavelength)
10GHz=3cm, 60 GHz=0.5cm and so on. This is a point where the chip dimensions become compatible with the wavelength and momentum simulations are mandatory. Also in addition to the structural simulations, you need to treat everything as a transmission line (or in other words, lumped parameter approach doesn't work) every component is modeled as a distributed entity.
Needless to say, these are separate fields of study.
Part of the terminology is historical. Integrated circuits were limited in frequency range and so hybrid modules were used for microwave frequencies. These were called microwave integrated circuits or MIC for short. This was a bit of advertising hot air to us the term integrated circuit. This is sort of like the VW microbus which was not 1e-6 the size of a regular bus but about 1/3 the size. By using a hybrid, good components like quality capacitors and high ft transistors could be used. These were expensive and only used in military applications and some commercial ones where their size and performance justified the cost.
Once processing technology improved, the microwave circuits were completely integrated and they were called monolithic microwave integrated circuits or MMIC for short.
The term RFIC was another bit of advertising hot air to differentiate a new market segment very much like using the term wireless instead of radio.
Each of these technologies are for different markets and are different processing technologies as discussed above.
Advertising hot air is very common. When women's suits with pants on the bottom became popular 40 years ago they were called pantsuits. They became unpopular and about 10 years ago they became popular again but were called trousersuits even though they were identical to pantsuits.
Momentum is an electromagnetic simulation tool. You can read more about it on the HP website.(currently Agilent) Momentum runs, for example, EM simulations to extract parameters for an inductor using Green's functions. Since i am not a CAD person, this is as far as i would go. it can also run some structural simulations but not really designed for that.
I was a little vague, I apologize. When i say structural simulation, i mean, hp's HFSS (high frequency structural simulator) or the like. This would be an essential part of designing microstrip lines or waveguides for MMIC design. other tools can be CST Microwave Studio, FEMLAB, Genesys Suite, Microwave Office that have some if not all the functionality.
For terminology, MMIC refers to "millimeter wave" integrated circuits as far as i know. At least, this is what is meant by MMIC currently.
and anything above that is usually referred to as being in the MMIC domain. The division is sort of fuzzy, since you can have 40GHz circuits on cmos these days.
Just to clarify few things that uncle_urfi statements:
What he meant by anything above that is really anything above that frequency in this case 30GHz or in wavelength it is 1cm (lambda). This also means anything below the wavelength of 1 cm will constitute an MMIC. (f=c/lambda)