[Show abstract][Hide abstract] ABSTRACT: A 1.95-GHz asymmetric multilevel outphasing (AMO) transmitter with class-E GaN power amplifiers (PAs) and discrete supply modulators is presented. AMO transmitters achieve improved efficiency over envelope tracking (ET) transmitters by replacing the continuous supply modulator with a discrete supply modulator implemented with a fast digital switching network. Outphasing modulation is used to provide the required fine output envelope control. A 4-level supply modulator is implemented that allows for fast and efficient discrete envelope modulation with up to 28-V supply voltages using low-voltage gate drivers and time-alignment logic. With two class-E GaN PAs that achieve 62.5% power-added efficiency (PAE) at 40- dBm peak output power, the AMO transmitter delivers 42.6- dBm peak output power at 1.95-GHz. For a 16-QAM signal at 36-dBm output power, the transmitter achieves 44.2/42.8/41.4% average system efficiency and 2.0/2.1/3.1% EVM for 10/20/40-MHz channel bandwidth, respectively.
[Show abstract][Hide abstract] ABSTRACT: We present a high-efficiency transmitter based on asymmetric multilevel outphasing (AMO). AMO transmitters improve their efficiency over LINC (linear amplification using nonlinear components) transmitters by switching the output envelopes of the power amplifiers among a discrete set of levels. This minimizes the occurrence of large outphasing angles, reducing the energy lost in the power combiner. We demonstrate this concept with a 2.5-GHz, 20-dBm peak output power transmitter using 2-level AMO designed in a 65-nm CMOS process. To the authors' knowledge, this IC is the first integrated implementation of the AMO concept. At peak output power, the measured power-added efficiency is 27.8%. For a 16-QAM signal with 6.1dB peak-to-average power ratio, the AMO prototype improves the average efficiency from 4.7% to 10.0% compared to the standard LINC system.
Power Amplifiers for Wireless and Radio Applications (PAWR), 2011 IEEE Topical Conference on; 02/2011
[Show abstract][Hide abstract] ABSTRACT: form only given, as follows. A low-power, low-complexity adaptive digital predistortion system is presented for linearizing WLAN PAs. Second-order quadrature delta-sigma modulation is employed with phase alignment and gain normalization for maintaining modulator stability. An experimental system linearizes a 2.4-GHz PA transmitting a 20.2-dBm WLAN signal with 10-MHz bandwidth. After 523-millisecond background adaptation, the experimental predistortion system improves the EVM of the unlinearized PA from 5.90% to 0.96%.
IEEE MTT-S International Microwave Symposium digest. IEEE MTT-S International Microwave Symposium 01/2011; DOI:10.1109/MWSYM.2011.5973438
[Show abstract][Hide abstract] ABSTRACT: A digital-to-RF phase modulator based on a single current-steering DAC is presented, including a carrier pre-rotation scheme that prevents phase inaccuracy due to carrier feedthrough. The phase modulator has been fabricated in a standard 65-nm CMOS process and draws 1.9 mW from a 1-V supply. The modulator achieves 12-bit resolution at a measured 200 MSamples/second, state-of-the-art performance in both resolution and sampling speed. It has sufficient speed to allow for oversampling to shape the output spectrum and therefore reduce filtering requirements, as demonstrated through a 32x oversampled 8-PSK signal at 6.25 MSymbols/second with under 6.1% EVM.
Radio Frequency Integrated Circuits Symposium (RFIC), 2011 IEEE; 01/2011
[Show abstract][Hide abstract] ABSTRACT: Since its invention in 1922, the super-regenerative amplifier (SRA) has been used in a variety of short-range, low-power, and/or low-cost wireless systems due to its simple implementation and excellent performance for a given power budget. Growing demand for ultralow-power receivers for short-range radios has recently reawakened an interest in the theory and design of SRAs. Building on recent work and using reasonable assumptions and approximations, we present a frequency-domain model for analyzing SRAs. We then use these models to predict the response of an SRA to arbitrary deterministic and stochastic signals including sinusoids, pulsed-sinusoids, and additive white Gaussian noise. Using the results, we present formulas for calculating the sensitivity and selectivity of SRAs. We also introduce the concept of a trigger-time that is particularly useful for accurately determining the optimal threshold in on-off keying (OOK) receivers and helps avoid the problems introduced by nonlinearity in SRAs. Finally, we present a prototype OOK SRA that achieves a sensitivity of -90 dBm at a bit rate of 300 kbps (BER of 10<sup>-3</sup>) while consuming 500 ??W, and show that its measured sensitivity matches theory within 1 dB.
IEEE Transactions on Microwave Theory and Techniques 10/2010; 57(12-57):2882 - 2894. DOI:10.1109/TMTT.2009.2033843 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: We present the first known energy management IC to allow low-power systems, such as biomedical implants, to optimally use ultracapacitors instead of batteries as their chief energy storage elements. The IC, fabricated in a 0.18 μm CMOS process, consists of a switched-capacitor DC-DC converter, a 4 nW bandgap voltage reference, a high-efficiency rectifier to allow wireless recharging of the capacitor bank, and a switch matrix and digital control circuitry to govern the stacking and unstacking of the ultracapacitors. The stacking procedure allows for more than 98% of the initial energy stored in the capacitors to be removed before the output voltage drops unsuitably low. The DC-DC converter achieves a peak efficiency of 51% for loads between 10 and 100 μW, operates for input voltages between 1.25 and 2.5 V.
[Show abstract][Hide abstract] ABSTRACT: We present a high-efficiency transmitter architecture based on asymmetric multilevel outphasing (AMO), but with a new method of generating discrete amplitude levels from the constituent amplifiers. AMO and multilevel LINC (ML-LINC) transmitters improve their efficiency over LINC by switching the supplies of the power amplifiers (PAs) among a discrete set of voltages. This allows them to minimize the occurrence of large outphasing angles. However, it is also possible to generate a discrete set of amplitudes by varying the duty cycle of the waveform that drives the PAs. The chief advantage of this discrete pulse width modulation (DPWM) is hardware simplicity, as it eliminates the need for a fast, low-loss switching network and a selection of power supply voltages. We demonstrate this concept with a 48-MHz, 20-W peak output power AMO transmitter using a four-level DPWM. At peak output power, the measured power-added efficiency is 77.7%. For a 16-QAM signal with a 6.5-dB peak-to-average power ratio, the AMO prototype improves the average efficiency from 17.1% to 36.5% compared to the standard LINC system.
[Show abstract][Hide abstract] ABSTRACT: We describe a new outphasing transmitter architecture in which the supply voltage for each PA can switch among multiple levels. It is based on a new asymmetric multilevel outphasing (AMO) modulation technique which increases overall efficiency over a much wider output power range than the standard LINC system while maintaining high linearity. For demonstration, the overall transmitter is simulated in a 65 nm CMOS process with HSUPA and WLAN signals. The simulation results show an efficiency improvement from 17.7% to 40.7% for HSUPA at 25.3 dBm output power and from 11.3% to 35.5% for WLAN 802.11g at 22.8 dBm while still meeting system linearity requirements.
Radio Frequency Integrated Circuits Symposium, 2009. RFIC 2009. IEEE; 07/2009
[Show abstract][Hide abstract] ABSTRACT: We demonstrate energy-efficient low-complexity adaptive linearization for wideband handset power amplifiers (PAs). Due to power overhead and complexity, traditional wideband linearization techniques such as adaptive digital predistortion (DPD) thus far have not been used for wideband handset transmitters. Our energy-efficient lookup table training strategy resulted in a training energy of 1.83 nJ/entry for a 5-MHz bandwidth WiMAX orthogonal frequency division multiple access (OFDMA) transmission, which represents more than 40times improvement over state-of-the-art DPD implementations. Our experimental prototype transmitter achieves a maximum of 9.9-dB improvement of adjacent channel leakage power at 5.15-MHz offset with 22.0-dBm channel power in the 5-MHz bandwidth WiMAX-OFDMA transmission. This linearity improvement offers 26.5% savings in PA power consumption by reducing power backoff.
IEEE Transactions on Microwave Theory and Techniques 11/2008; 56(10-56):2248 - 2258. DOI:10.1109/TMTT.2008.2003139 · 2.24 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: We present a new adaptive power amplifier (PA) linearization technique. We leverage analog Cartesian feedback (CFB) to train a Cartesian look-up table, reducing DSP and power amplifier modeling requirements to a minimum and eliminating model convergence as a design issue. Because the CFB system does not continuously operate, we overcome the bandwidth limitation traditionally associated with this technique. In addition, we exploit sample averaging to greatly relax the noise requirements of the analog feedback path. We implemented a prototype 900-MHz direct-conversion transmitter with a class-A PA. We measured a 10-dB reduction of out-of-band distortion products with no noise floor degradation for 40-MHz-bandwidth, 16-QAM test signals.
[Show abstract][Hide abstract] ABSTRACT: Sponsorship: Texas Instruments, DARPA Biomedical sensors are used to measure a myriad of biopotential signals including electroencephalogram (EEG), electrocardiogram (EKG), electromyogram (EMG), and neural field potential (NFP) signals , , . Most of the useful information in these signals resides in the frequency range of 0.5 Hz to 1 kHz, allowing ultra-low power circuits to be used when processing them. This is critical for systems that are implanted, since energy is extremely scarce, and the lifetime of the device must be on the order of 10 years. Unfortunately, these signals are often as small as 10 µVs, and their low frequency location make them vulnerable to aggressors such as DC offset, powerline noise, and flicker noise. DC offset can result from charge accumulation at the interface between the metal electrodes and the skin, and also from amplifier offsets caused by random mismatches. While chopper stabilization has proved effective at mitigating the effects of amplifier DC offset and flicker noise, electrode DC offset cannot be removed through chopping and must be high-pass filtered at the front end of the system to prevent saturation , , . Powerline noise, typically at 50 or 60 Hz, is mostly a common-mode signal that requires adequate common-mode rejection. However, if there are mismatches or inductive loops in the electrodes, these aggressors can become differential-mode signals, corrupting the desired signal, and potentially saturating the system. In closed-loop deep brain stimulation systems, another aggressor arises from stimulation artifacts . In that case, the NFPs can be much smaller than stimulation artifacts placing stringent requirements on the dynamic range of the system and potentially leading to signal corruption. We propose a mixed-signal sensor interface that mitigates the effects of all of the aforementioned aggressors in an area efficient manner. Area efficiency is particularly compelling in implantable devices that use tens or hundreds of electrodes, such as neural recording systems . The proposed system, shown in Figure 1, uses a chopper stabilized operational amplifier with capacitive feedback to achieve accurate gain (The system is shown as single-ended for simplicity, but is implemented in a fully differential manner). Figure 2 shows a simplified schematic of the amplifier, including a novel input chopper that creates a switched capacitor resistance between its inputs and a reference voltage. This resistance is shown as Rp in Figure 1 and is used to create a high-pass filter with a corner frequency well below 1 Hz, while setting the common-mode voltage of the input to a desired level. The pole frequency is actually set by the Miller-multiplied feedback capacitor Cf and is inversely proportional to the amplifier's gain A V , allowing a reduction of many orders of magnitude in component sizes. An additional feedback path is introduced that includes the filter, ADC, DSP, and a feedback DAC. This path can be used to notch out unwanted signals such as powerline noise or stimulation artifacts before they can saturate the system.
[Show abstract][Hide abstract] ABSTRACT: Sponsorship: The Lemelson Foundation Presidential Fellowship Until recently, few medical implantable devices existed and fewer still provided the capability for wireless transmission of informa-tion. Most devices capable of data transmission did so through inductive coupling, which requires physical contact with the base-station and allows for only low data rates. In 1999, the FCC created the Medical Implant Communications Service (MICS) band in the range of 402–405 MHz specifically for medical te-lemetry . The MICS band plan allows for RF communication between a medical implant and a base-station that is up to two meters away. This research seeks to design a transceiver specifi-cally optimized for low-power, short-distance data transmission in a temperature-regulated environment, i.e., the human body. We do this by pushing as much complexity as possible out of the implant and into the base-station, taking advantage of the attri-butes of the environment, such as temperature control and slow transients, and incorporating the antenna into the oscillator for reduced power and improved performance. By optimizing the transceiver for reduced volume and power, we hope to extend the battery lifetime and functionality of medical implants for greater comfort and benefits to patients. Figure 1a shows a conventional direct up-conversion transmit-ter that comprises a digital baseband, digital-to-analog convert-ers (DAC), low-pass filters, up-conversion mixers, an I/Q phase generator, a power amplifier (PA), a frequency synthesizer, and an antenna. We propose a much simpler, almost all-digital imple-mentation (see Figure 1b, composed of a digital baseband and a digitally controlled oscillator (DCO)). Instead of direct I/Q up-conversion, we propose using minimum frequency-shift keying to directly modulate the DCO with baseband information. We ex-ploit the inherent temperature regulation of the human body and the lax frequency stability requirements of the MICS standard to replace the frequency synthesizer with a much slower frequency-control loop, which incorporates the base-station. Furthermore, we create a linear digital-to-frequency converter by using pre-distorted capacitor banks for coarse and fine frequency tuning. Instead of driving the antenna with a matched PA, we exploit the low radiation power requirement to incorporate a loop antenna into the DCO. The inherently high Q of the antenna leads to im-proved noise performance for a given amount of power. Figure 2 shows the differential Clapp DCO including coarse-and fine-tuning capacitor banks, a circuit model for the loop antenna, and a cross-coupled pair of transistors for power reduction . Figure 1: (a) Conventional direct up-conversion transmitter. (b) Proposed transmitter comprised of a digital baseband, a digitally controlled oscillator, and a loop antenna.