High Voltage Level-Shifter Circuit Design for Efficiently ...

1 high Voltage Level-Shifter Circuit Design for EfficientlyHigh Voltage Transducer DrivingHao-Yen TangElectrical Engineering and Computer SciencesUniversity of California at BerkeleyTechnical Report No. UCB/EECS-2014-203 1, 2014 Copyright 2014, by the author(s).All rights reserved. Permission to make digital or hard copies of all or part of this work forpersonal or classroom use is granted without fee provided that copies arenot made or distributed for profit or commercial advantage and thatcopies bear this notice and the full citation on the first page. To copyotherwise, to republish, to post on servers or to redistribute to lists,requires prior specific permission.

2 Acknowledgement The authors would like to thank the Taiwan SemiconductorManufacturing Company (TSMC) for chip fabrication andPramod Murali for helpful Voltage Level-Shifter Circuit Design forEfficiently high Voltage Transducers DrivingHao-Yen TangBerkeley Sensor & Actuator Center,University of California, Berkeley, CA : This report describes a level shifter with crow-bar current suppression is proposed to achieve simultaneouslyhigh switching speed and low power dissipation for high -voltagetransducer driving. Unlike prior implementations, the new circuitdoes not require additional high - Voltage supplies nor static powerconsumption.

3 The prototype switches up to 32 V with V input consuming less than pJ/V2per transition. Delaysas low as 16 ns and ns for falling and rising transitionsrespectively with V input are ideal for time-critical phasedarray beamforming applications and could be further reducedto ns and ns when driven with V supply. Togetherwith buffer circuitry, the Circuit is able to handle loads up to100 pF Efficiently with frequency above 10 MHz, which makesthe Circuit favorable for high Voltage transducers application suchas ultrasonic INTRODUCTIONMany sensor interfaces.

4 Need high Voltage actuation. Appli-cations include electrostatic actuation or ultrasonic transducers[1], [2], where typical operating Voltage varies from above10V for piezo-electric transducers and Piezo-electric Micro-machined Ultrasonic Transducers (PMUT) to above 50V forCapacitive Micro-machined Ultrasonic Transducers(CMUT).Furthermore, the high resolution requirements necessitiesabove 10 MHz driving frequency for the transducers with over10pF parasitic capacitance. To address the requirements weare presenting a driver capable of both high Voltage and highband-width remaining part of this report is organized as following:Section II.

5 Briefly describes the high Voltage IC process andtypical high - Voltage sensor interface Circuit . Section III. sum-marizes previous solutions for high - Voltage driving. SectionIV. explains the Circuit we proposed for efficient high -voltagelevel-shifting. Experiment results are given in Section V, andfinally conclusion is in Section are several high - Voltage processes available from var-ious foundries. For example, the TSMC 32 ,consists of normal transistors operating at nominalVDD= ,and special-designed high - Voltage transducers operating at arelatively high supply voltageHV VDD=32V.

6 Due to potentiallatch-up problems, the high - Voltage transistors is inside high - Voltage wells isolated with high - Voltage guard-ring and high - Voltage bottom isolation layer. Fig. 1 gives an example forThe high - Voltage MOSFET (HVMOS) has thicker oxidethickness to prevent oxide breakdown from high gate clear drawback is reducedCoxand turn-on current. Further-Fig. 1. Typical layout for high - Voltage process (a)topview (b) , the minimum gate length of HVMOS in this process islimited to m. Also HV devices need to be spaced furtherthan LV transisters, not to mention area consumed by guard-ring itself.

7 These several drawback make it challenging todesign a high -speed high Voltage Circuit Efficiently withoutlarge area or power minimize the area and power overhead, typical high - Voltage sensor interface circuits perform most of the signalprocessing such as delay and frequency control in low voltagedomain, following by a Level-Shifter to shift up the signal tohigh- Voltage domain as illustrated in Fig. 2. The level-shifteroften consumes significant power due to high supply voltageand has limited bandwidth due to the lowfTof thick-oxidehigh Voltage transistors.

8 Level-Shifter described in this thesesavoids these PREVIOUS WORK SUMMARYT here are several existing solution to convert low-voltagecontrol signal to high Voltage domains. Due to poor transcon-ductance of high - Voltage transistors, straightforward solutionsbased on linear amplifiers suffer from low efficiency andhigh Circuit complexity. Similar to PWM audio amplifiers [3],drivers with only two output levels avoid these solutions exploit the inherent low-pass characteristicof the transducer to filter out harmonics but require highFig. 2.

9 Typical high - Voltage sensor interface speed to meet bandwidth requirements and meetsignal-to-noise ratio PWM driver interfaces the low- Voltage digital controlinput to a high Voltage switched output. Depending on theapplication, output voltages in excess of 30 V are required,corresponding to step-up ratios of up to 16 with V , this entire gain is realized in a single stage to minimizecircuit complexity and power in intermediate 3a shows an implementation of a latch-based Level-Shifter comprising an NMOS differential pair with low-voltageinput and a PMOS negative resistance load [4].

10 Althoughsimple, this Circuit has several drawbacks. Firstly, the largeoverdrive Voltage of the PMOS devices set by the high - Voltage supply necessitates very wide NMOS transistors tooverpower the load with the low- Voltage input signal. Theconsequent high input capacitance limits the maximum op-erating speed and increases the power dissipation in thedriver Circuit . Secondly, although static, the Circuit is subjectto high dynamic power dissipation, owing to the crow-barcurrent flowing during the transition when both the NMOS andPMOS transistors are on, as indicated in the timing for step-up ratios in excess of five, high operatingspeeds and for small capacitive loads characteristics that arerepresentative of many MEMS applications [1] these lossescan significantly degrade the solutions have been proposed to overcome theselimitations.