A 120GHz 10Gb/s phase-modulating transmitter in 65nm LP CMOS

Abstract

This paper presents a 120GHz fully integrated 65nm low power (LP) CMOS transmitter which achieves data rates above 10Gb/s. At these high frequencies an extremely high bandwidth is available. This allows multi gigabit per second communication which provides an answer to the ever increasing demand for higher data rates in wireless systems. However, wideband modulation of a 120GHz signal in 65nm LP CMOS is a challenge. To achieve a high data rate at these high frequencies, a straightforward and high speed phase modulator architecture is implemented. Figure 16.7.1 shows the complete schematic of the transmitter. The LO input signal is first buffered by a 2-stage buffer amplifier. The phase generation is done by a novel differential branch line coupler that generates two differential quadrature signals. These 2 signals are fed into two 6-stage driver amplifiers. By means of differential splitters, each signal is split in an in-phase and an inverted signal. This results in 4 differential signals with relative phases of 0, 90, 180 and 270 degrees, allowing phase modulation. The actual modulation is carried out by a 4-to-1 analog multiplexer which selects the generated 120GHz phases depending on the modulation scheme. The high-speed MUX is directly driven by digital signals and takes over the function of a conventional DAC and upconverter. The output of the MUX is buffered and fed into a 9-stage differential PA. The novel differential branchline coupler that generates the four phases, is based on a single ended branchline coupler. A 120GHz on chip signal has a wavelength of about 1.2mm. Conventional branchline couplers have a width and length of about /4, which is thus about 300µm. The use of slow-wave transmission lines, which have a higher artificial relative permittivity, resulting in lower phase velocity and shorter wavelengths, leads to a very compact design. The differential branchline coupler consumes an area of 140µm by 180µm which is almost 4 times less than a conventional branchline coupler. The 50 differential lines of the coupler are designed in the top metal layers and only have a bottom shield. A branchline coupler also needs a transmission line with a value of 50 / 2. To implement this 35 differential slow-wave line, an extra top shield is added. The use of a differential structure results in a less complicated and more compact slow-wave line than the single ended version. On top of that, the inverted signals are also available. The measured phases that are generated by this transmitter are shown in figure 16.7.3. The analog multiplexer consists of 4 differential pairs which can be switched on or off by their tail current source (See figure 16.7.2). The tail current source is directly driven by a buffered differential digital signal. If two quadrature phase channels are switched on together, an additional 45° interpolated phase can be realized, resulting in an 8-point QAM constellation. The AM of this signal can be removed by pushing the PA into saturation, which also improves the overall power efficiency ...

A plastic waveguide receiver in 40nm CMOS with on-chip bondwire antenna

“…The power that comes out of the metal waveguide is measured with an Erickson PM4 power meter. Unless stated otherwise, this level is set to 0 dBm, a value that is certainly feasible to be generated with modern CMOS processes [2], [3]. To demonstrate multilevel ASK demodulation, the AWG was used to generate a four level ASK signal.…”

Energy-efficient adaptive wireless NoCs architecture

“…4: Power amplifier trends in integrated transmitters implemented using compound (III-V) and silicon-based (SiGe HBT and CMOS) devices. Data collected from [17], [18], [19], [20] of carrier/modulation frequency. SiGe HBTs are more suitable for WiNoCs due to their power levels and material engineering techniques on silicon bipolar transistors compared to high performance III-V HEMTs with poor integration potential.…”

CMOS mm-wave transceiver techniques beyond 50 GHz

“…Design of RF front-ends at mm-wave carriers (100 GHz) and beyond is a relatively new topic with most emphasis placed on the circuit design aspects. One major challenge of mm-wave front-ends operating at frequencies beyond 100 GHz is that the stage gain for power amplifiers is quite low [1,2,3], typically only 3-6 dB of gain per stage. This low available gain makes the front-end extremely sensitive to the effects of process variation as a small change in the CMOS device parameters may drastically reduce the front-end performance.…”