Sub-threshold operation and cross-hierarchy design for ultra low power wearable sensors

Calhoun, Benton H.; Bolus, Jonathan F.; Khanna, Sudhanshu; Jurik, Andrew D.; Weaver, Alfred C.; Blalock, Travis N.

doi:10.1109/iscas.2009.5118036

“…With advances in Internet of Thing (IoT) applications and the expansion of mobile devices, energy consumption has become a primary focus of attention in integrated circuits design [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18]. While IoT applications cover a broad range of products from wearable devices, smart houses, automotive devices, smart meters to inspection tools and many others, more than 50% of the market is dominated by battery operated devices.…”

Section: Introductionmentioning

confidence: 99%

0.45 v and 18 μA/MHz MCU SOC with Advanced Adaptive Dynamic Voltage Control (ADVC)

Zangi¹,

Feldman²,

Hadas³

et al. 2018

JLPEA

View full text Add to dashboard Cite

An ultra-low-power MicroController Unit System-on-Chip (MCU SOC) is described with integrated DC to DC power management and Adaptive Dynamic Voltage Control (ADVC) mechanism. The SOC, designed and fabricated in a 40 nm ULP standard CMOS technology, includes the complete Synopsys ARC EM5D core MCU, featuring a full set of DSP instructions and minimizing energy consumption at a wide range of frequencies: 312 K-80 MHz. A number of unique low voltage digital libraries, comprising of approximately 300 logic cells and sequential elements, were used for the MCU SOC design. On-die silicon sensors were utilized to continuously change the operating voltage to optimize power/performance for a given frequency and environmental conditions, and also to resolve yield and life time problems, while operating at low voltages. A First Fail (FFail) mechanism, which can be digitally and linearly controlled with up to 8 bits, detects the failing SOC voltage at a given frequency. The core operates between 0.45-1.1 V volts with a direct battery connection for an input voltage of 1.6-3.6 V. Measurement results show that the peak energy efficiency is 18µW/MHz. A comparison to state-of-the-art commercial SOCs is presented, showing a 3-5× improved current/DMIPS (Dhrystone Million Instructions per second) compared to the next best chip.

show abstract

“…In extreme cases, the combination of process and temperature variations can even cause certain circuits to malfunction. Therefore, specific measures need to be taken to deal with this drawback.Sub-threshold and near-threshold circuits have received a lot of attention in the research community during the last decade [6-8,10-13] with successful ultra-low voltage chip implementations: microprocessors [6,8,[10][11][12], as well as dedicated ASICs for biomedical applications [7,13,[33][34][35][36], for wireless sensor nodes [10,15,17,[37][38][39], communication [40,41], image processing [42,43], RFIDs [44] and many others. While previous research indicated significant power reduction and successful minimum energy operation, it did not present complete MCU SoC solutions that are suitable for mass production.This paper presents a 40 nm MCU System-on-a-Chip (SoC), which includes the complete Synopsys ARC EM5D core MCU and features a full set of DSP instructions [29].…”

mentioning

confidence: 99%

0.45v and 18pA/MHz MCU SOC with advanced Adaptive Dynamic Voltage Control (ADVC)

Zangi¹,

Feldman²,

Shor

³

et al. 2017

2017 IEEE SOI-3D-Subthreshold Microelectronics Technology Unified Conference (S3S)

3

0

View full text Add to dashboard Cite

An ultra-low-power MicroController Unit System-on-Chip (MCU SOC) is described with integrated DC to DC power management and Adaptive Dynamic Voltage Control (ADVC) mechanism. The SOC, designed and fabricated in a 40 nm ULP standard CMOS technology, includes the complete Synopsys ARC EM5D core MCU, featuring a full set of DSP instructions and minimizing energy consumption at a wide range of frequencies: 312 K-80 MHz. A number of unique low voltage digital libraries, comprising of approximately 300 logic cells and sequential elements, were used for the MCU SOC design. On-die silicon sensors were utilized to continuously change the operating voltage to optimize power/performance for a given frequency and environmental conditions, and also to resolve yield and life time problems, while operating at low voltages. A First Fail (FFail) mechanism, which can be digitally and linearly controlled with up to 8 bits, detects the failing SOC voltage at a given frequency. The core operates between 0.45-1.1 V volts with a direct battery connection for an input voltage of 1.6-3.6 V. Measurement results show that the peak energy efficiency is 18µW/MHz. A comparison to state-of-the-art commercial SOCs is presented, showing a 3-5× improved current/DMIPS (Dhrystone Million Instructions per second) compared to the next best chip. efficient) Systems on a Chips (SoC) that can support both minimum energy operation, and also reliably adapt their operating voltage to different environmental conditions is very challenging.Unfortunately, conventional SoCs, using supply voltages in the range of 0.9-1.8 V, are not suitable for IoT applications. In these SoCs, the "on" transistors operate in "super-threshold" region, far above the switching threshold of a transistor. In this region, the "Ion" current of a transistor is very strong, which results in a ratio of many orders of magnitude between the "Ion" and the "Ioff" (parasitic leakage) currents. This results in a very fast and reliable, albeit power hungry operation.Low voltage operation in the "sub-threshold" or "near-threshold" regions have been found to be advantageous in dramatically reducing energy dissipation [16,18,19]. However, the power supply reduction is accompanied by a number of problems and significant challenges, especially in mass production designs [22][23][24][25][26]. The low voltage associated with frequency reduction is not suitable for all modes of operation, and an adaptive voltage control mechanism is required. Lower supply voltages also mean lower noise margins, reduced yield, and increased vulnerability to process variations and temperature fluctuations [27][28][29][30][31][32]. The characteristics of semiconductor behavior in sub/near-threshold are not well represented by standard transistor models and are different from those in the super-threshold region, resulting in different sizing and ratio optimizations.The subthreshold current is exponentially dependent on the transistor's threshold voltage. Hence, process variations that substantially affect the threshold vol...

show abstract

“…For example, a wireless ECG monitor could derive heart rate, or detect instances of cardiac arrhythmia, rather than transmitting the complete ECG signal. This often has the effect of reducing the total power consumption, because wireless data transmission is frequently the dominant consumer of power in a wireless sensor, and the energy expended to perform feature detection and compression can frequently be made less than the energy saved by reducing the amount of data to be transmitted [24] [25].…”

Section: Wireless Sensor Architecturementioning

confidence: 99%

Design of mm-Sized Wirelessly Powered Sensors

Bolus

¹

Self Cite

0

View full text Add to dashboard Cite

This work presents several circuits that can be used to improve the range and decrease the size of millimeter-sized wirelessly powered sensors. It is shown that this can be accomplished by decreasing the power consumption of the sensor, and several circuits common to wireless sensors are presented that consume very low amounts of power, in the 100 nW range. A wireless power transmission system that can generate power in the range of 1 µW is also presented. In completing my doctorate I relied, in no small part, on the support of many colleagues, family and friends. Without their support the work presented here would not have been possible, and I am very grateful for their efforts. A few of these people in particular I would like to distinguish here. First, I would like to thank my advisor, Travis Blalock, for his guidance and support. He allowed me the opportunity to work on a wide variety of interesting and challenging projects during my time at UVa, and was an invaluable help during all stages of my research. He always displayed great confidence in my abilities, and was always receptive and encouraging of my ideas. I would also like to thank the my doctoral committee, Ben Calhoun, Alf Weaver, Ron Williams and Scott Barker, for all their efforts and helpful feedback. In particular I would like to thank my committee chair, Ben Calhoun. He provided me with space on several chips to implement my designs, offered substantial constructive feedback on a great deal of my writing, and was always highly motivating. Many of my fellow students are also deserving of my thanks. Stuart Wooters was always available to help with any difficulties I was having, and our conversations about research proved very useful to my work. He also constructed the pad ring for the demodulator and wireless power transmission chips used in Chapters 3 and 4. Yanqing Zhang helped me with the synthesis of some of the digital logic used for iii the demodulator. Kyle Alexander Craig constructed the pad ring for the random identification chip described in Chapter 2. I would also like to thank Steve Jocke, Andrew Jurik, and Peter Beshay for their assistance. Many members of my family are deserving of my thanks. In particular, I would like to thank my grandfather, Fred Wikner, who has always maintained a great interest in my education, and been a strong motivating force in the pursuit of my doctorate. Lastly, but most importantly, I would like to thank my parents John and Andrea, and my brother Christian, for their unwavering love and support. They have always encouraged my interest in technical things, and have provided me with everything I needed to pursue my education. Their confidence in me has been invaluable. I cannot imagine a more supportive family, and for this I will always be grateful.

show abstract

Sub-threshold operation and cross-hierarchy design for ultra low power wearable sensors

Cited by 9 publications

References 8 publications

0.45 v and 18 μA/MHz MCU SOC with Advanced Adaptive Dynamic Voltage Control (ADVC)

0.45 v and 18 μA/MHz MCU SOC with Advanced Adaptive Dynamic Voltage Control (ADVC)

0.45v and 18pA/MHz MCU SOC with advanced Adaptive Dynamic Voltage Control (ADVC)

Design of mm-Sized Wirelessly Powered Sensors

Contact Info

Product

Resources

About