An Accurate BJT-Based CMOS Temperature Sensor With Duty-Cycle-Modulated Output

Abstract

This paper describes the design of a precision BJT-based temperature sensor implemented in standard 0.7µm CMOS technology. It employs substrate PNPs as sensing elements, which makes it insensitive to the effects of mechanical (packaging) stress and facilitates the use of low-cost packaging technologies. The sensor outputs a duty-cycle-modulated signal, which can easily be interfaced to the digital world and, after low-pass filtering, to the analog world. In order to eliminate errors caused by component mismatch, chopping and Dynamic Element Matching (DEM) techniques have been applied. The required component shuffling was done concurrently rather than sequentially, resulting in a fast DEM scheme that saves energy without degrading accuracy. After a single-temperature trim, the sensor's inaccuracy is ±0.1ºC (-20ºC to 60ºC) and ±0.3ºC (-45ºC to 130ºC), respectively. Measurements of sensors in different packages show that the package-induced shift is less than 0.1ºC. Measurements of 8 sensors over 367 days show that their output drift is less than 6mK. While dissipating only 200µW, the sensor achieves a resolution of 3mK (rms) in a 1.8ms measurement time, and a state-of-the-art resolution Figure of Merit (FoM) of 3.2pJK 2. This combination of high accuracy, high resolution, high speed and low energy consumption makes this sensor suited for commercial and industrial applications.

A BJT-Based Temperature-to-Digital Converter With ±60 mK ( $3~\sigma$ ) Inaccuracy From −55 °C to +125 °C in 0.16-μm CMOS

“…In previous TDCs [6]- [9], [14]- [16], different techniques have been used to reduce the effects of spread in the various parameters of (4) and (5). Choosing a small biasing current (I E ) mitigates the effect of r S (i.e., the last terms in (4) and (5)) and its variation.…”

“…In order to increase the accuracy of ∆V BE , dynamic element matching (DEM) has been used [6]- [9], [14]- [16] to accurately define p E (first term in (5)). This will improve the accuracy of ∆V BE , provided that β is current independent (i.e., ∆β = 0, in the second term of (5)), or that β This is the author's version of an article that has been published in this journal.…”

“…To mitigate such errors, the current mirror ratio p Eb is dynamically matched (DEM1), in the same way as the current mirror ratio p E (DEM2). For the same reasons, a similar approach was recently used in [16]. Furthermore, to ensure that biasing currents are accurately copied from the biasing-circuit to the bipolar-core, the two banks of current mirrors are dynamically swapped (Bank-Swap).…”

CMOS based thermal detector for processor

“…Due to the high-power usage of the processor, it is even challenging to add-on the thermal sensor into the processor [6]. For an on-chip temperature sensors, it must be compatible with the technology available in the process, maintain a small area, has good accuracy with low power consumption over standard process variations and over typical supply voltage variations [1,4,[6][7].…”

Unusual Operation of the Junction Transistor Based on Dynamical Behavior of Impurities

“…It was for three decades the active device of choice in the analog design of integrated circuits used in wide applicability in electronics, including computers, televisions, mobile phones, audio amplifiers, industrial control, and radio-frequency (RF) circuits of the wireless systems [1], but today the use of the BJT has declined in favor of the CMOS technology in the design of digital integrated circuits (ICs) [2][3][4]. Due to the low impedance at the base, high transconductance, and output resistance compared to MOS devices, the BJT is characterized as currentcontrolled current source useful to compute nonlinear functions by their logarithm dependence between base-emitter voltage , collector current , and temperature, which recently has led to its integration into the CMOS-based architectures [5][6][7][8]. Currently, the BJT has been shown to be a viable option as transducer in electrochemical sensor, because it has significantly greater sensitivity and signal to noise ratio (SNR) and minor calibration requirements independent of voltages for operation, which implies that optimal measurements can be made over the entire sensing range, for example, to detect both [Cl − ] ions and biomolecules in portable diagnostics [9,10].…”