1. Sub-1V smart temperature sensors

Temperature-sensing RFID tags represent the next step in the RFID revolution. However, since they are battery or wirelessly powered, such sensors must dissipate very little power (< 1uW). BJTs are good candidates for this application, since they can achieve high (0.1°C) inaccuracy, and only require nanowatts of power. A proven way to achieve low power operation is to reduce the supply voltage, but a major challenge is the fact that VBE ~ 0.8V at -55°C. In this project, the use of MOSFETs instead of BJTs will be explored.

Partner/Location: NXP IC-Lab/TU Delft

K.A.A. Makinwa, “Smart Temperature Sensors in Standard CMOS,” (Proc. Eurosensors)
Procedia Engineering, pp. 930 – 939, Sept. 2010.

Essay topic: Micropower temperature sensors

2. Thermal diffusivity sensors for microprocessors

Modern microprocessors employ tens of temperature sensors to prevent hotspots and thermal failures. Such sensors should be compact, robust (digital output) and, most importantly, require minimal trimming. The last requirement cannot be met by traditional BJT-based sensors. However, a new class of sensors, called thermal diffusivity (TD) sensors, is much better in this regard. The goal of this project is to design a compact interface architecture for arrays of TD sensors.

Sponsor/Location: Texas Instruments/TU Delft

C.P.L. van Vroonhoven, M. Kashmiri and K.A.A. Makinwa, “CMOS Temperature Sensors
Based on Thermal Diffusion,” Proc. Therminic, pp. 140 – 143, Oct. 2009.

Essay topic: Temperature sensors for microprocessor thermal management

3. Solid-state airflow sensors

Server farms form the backbone of the internet. A major challenge is providing them with adequate and reliable cooling. The aim of this project is to develop a PCB-mounted solid-state wind sensor that can be manufactured in standard CMOS and with standard packaging.

Partner/Location: NXP/TU Delft

K.A.A. Makinwa and J.H. Huijsing, “A 2nd-order Thermal Sigma-Delta Modulator for flow
sensing,” Proc. IEEE Sensors, pp. 549 – 552, Oct. 2005.

Essay topic: Solid-state airflow sensors

4. High accuracy, low area front-end for magnetic sensors

Magnetic sensors are often employed in automotive applications for sensing the rotation speed or measuring the angular position of wheels and shafts. AMR (Anisotropic Magnetic Resistance) sensors, usually arranged in a Wheatstone bridge configuration, are often used in such applications. However, due to the cost pressure on automotive parts, the area of the readout electronics must be as small as possible. The goal of this project is to design a 3-channel readout ASIC for AMR sensors.

Partner/Location: TU Delft/NXP Eindhoven

M. Motz et al., “A chopped Hall sensor with small jitter and programmable “True Power-on”
function,” Journal of Solid-State Circuits, vol. 40, is. 7, pp. 1533 – 1540, 2005.

Essay topic: Fully-integrated magnetic sensors

5. Chopper amplifiers with low-bias current

A popular method of realizing amplifiers with micro-volt level offset is by the use of chopping. However, the residual offset and bias current of chopper amplifiers is often limited by the charge-injection associated with the switching operation of the chopper
switches. The resulting input currents can be quite large, in the order of several hundred pico-Amperes. The goal of this project is to explore various ways of reducing these input currents. Since this charge injection is a major source of residual offset, such techniques
should also lead to lower residual offset.

Sponsor/Location: TU Delft/ TU Delft

J. Xu et al., “Measurement and Analysis of Input Current Noise in Chopper Amplifiers,”
Proc. of ESSCIRC, Sept. 2012.

Y. Kusuda, “A 5.9nV/sqrt(Hz) Chopper Operational Amplifier with 0.78μV Maximum Offset
and 28.3nV/°C Offset Drift,” Digest ISSCC, Feb. 2011.

Essay topic: Electrometer amplifiers for high-impedance sensors

6. Micropower opamp with rail-to-rail input stage

This project involves the design of a micropower opamp with a supply current of about 10uA. The amplifier topology will be based on an existing design, to which a rail-to-rail stage will be added.

Partner: Maxim (Delft) / TU Delft

7. High-resolution Continuous-Time Sigma-Delta Converters

Most of the high-resolution (> 20-bit) sigma-delta ADCs reported in literature are based on switched-capacitor techniques. As a result, they employ relatively large sampling capacitors (kT/C noise), which must be driven by power hungry buffers. An alternative approach would be to use continuous-time techniques. Since the first integrator requires a current input, a transconductor is required to perform the initial voltage-to-current conversion. The challenge is then how to deal with the limited linearity of such a stage.

Partner: TU Delft/TU Delft

G. Singh et al., “A 20bit Continuous-Time Sigma Delta Modulator with a Gm-C Integrator,
120dB CMRR and 15 ppm INL,” Proc. of ESSCIRC, Sept. 2012.

D. Kim et al., “A continuous-time, jitter insensitive ΣΔ modulator using a digitally linearized
GmC integrator with embedded SC feedback DAC,”VLSI Circuits, pp. 38 – 39, June 2011.

Essay Topic: High Resolution (18-bit+) ADCs

8. High-resolution Low-Oversampling Sigma-Delta Converters

In many applications (audio, sensors, instrumentation, etc.) high resolution ADCs are required to convert analog signals into the digital domain. For resolutions over 16 bit, Sigma Delta ADCs have been a popular choice. Such ADCs use oversampling and noise shaping to move quantization errors out-of-band, thus increasing the dynamic range inband. Since amplitude resolution is traded for time resolution, a much high(er) sampling frequency is usually required, which is not always available. The aim of this project is to realize a Sigma Delta ADC that can provide high resolution at a limited clock frequency.

Partner/Location: TU Delft/NXP Eindhoven

Essay Topic: High Resolution (18-bit+) ADCs