IEEE SENSORS JOURNAL, VOL. 3, NO. 6, DECEMBER 2003 835 Constant Charge Operation of Capacitor Sensors Based on Switched-Current Circuits Rafael Nadal-Guardia, Anna Maria Brosa, and Alfons Dehé Abstract—A new method to perform the readout of capacitive sensors is presented in this paper. The sensitivity of integrated ca- pacitive sensors is investigated and compared considering two dif- ferent approaches. The traditional approach, in which a voltage source is connected via a resistor to the capacitor, and a completely different approach, in which the charge residing in the capacitor is kept constant. With this second technique, the charge injected to define the dc operating point is maintained constant and a linear relationship between the external excitation and the voltage moni- tored is obtained. As pointed out by the authors in previous works, the use of current pulses allows injecting and controlling the charge stored in an electrostatic microactuator. In order to avoid the ef- fect of the leakage currents, the charge is periodically refreshed. In this way, the charge is maintained constant inside the refresh cycle. In this paper, it is shown that this technique can be applied to capacitive sensors instead of electrostatic microactuators. The effects derived of using the temporal window, inside the refresh cycle to sense an external excitation, are further investigated. It is shown that the sensitivity of the sensor is increased in front of the obtained one with the voltage drive mode because the plates of the sensor can be fixed after the voltage pull-in distance without instability. Moreover, no voltage penalty is produced because the voltage developed across the transducer decreases after the pull-in distance. This method is introduced and studied through simula- tions, confirming the theoretical predicted behavior. Index Terms—Capacitive sensing, electrostatic transducers, post pull-in operation. I. INTRODUCTION S ILICON capacitive sensors have been extensively used thanks to many reasons like, for instance, the big amount of techniques available to do the readout [1] or the easy compatibility with standard CMOS fabrication processes [2]. For silicon capacitive sensors, switched-capacitor circuits using analog discrete-time approaches have been traditionally used. However, for some applications, like integrated capacitive microphones, easier implementations using a voltage source with a series capacitor connected to the sensor are enough to perform the signal readout [3]. The signal sensed is read Manuscript received December 7, 2001; revised July 15, 2003. This work was supported in part by the Bundesministerium Bildung und Forschung (BMBF) “Integrierte mikromechanische Si-Mikrofone für Array-Anwen- dungen (InFON)” V2089 Project, in part by Infineon Technologies AG, and in part by the FPI AP9677604835 program of the Spanish government. The associate editor coordinating the review of this paper and approving it for publication was Prof. Mona Zaghloul. R. Nadal-Guardia and A. M. Brosa are with the European Patent Office, Mu- nich, Germany. A. Dehé is with the Wireless Products, Infineon Technologies AG, Munich, Germany Digital Object Identifier 10.1109/JSEN.2003.820355 connecting the gate of a JFET or a MOSFET transistor in parallel with the sensor. The resistor and the voltage source used to bias the sensor define the power consumption and the sensitivity of the system. The use of the voltage source, giving all the charge demanded by the system when both electrodes are getting closer, avoids keeping the charge constant [4]. In this first scenario, the collapse of the moveable plate is produced for distances smaller than 2/3 of the gap distance [5], due to the well-known pull-in instability. However, such instability would be overcome if an “ideal” charge source would bias the transducer. No instability is defined in this charge drive mode when the parasitic capacitance in parallel with the transducer is smaller than one half of the initial capacitance of the transducer (that is at rest) [6]. Stable operation after the voltage pull-in distance can be achieved. Thus, the sensitivity can be easily increased. In this second scenario, a method to do the readout of capacitive sensors would fix first the charge and, afterwards, measure the voltage developed due to the external excitation. Several methods have been proposed to fix the moveable electrode of an electrostatically actuated structure after the pull-in distance [6]–[10]. The use of a series capacitor with the mechanical structure [6], [9] and the implementation of switched-capacitor techniques [7], are two of the most impor- tant method to stabilize the moveable electrode after the pull-in distance using a voltage source driving the transducer. However, problems related to the high voltage values required in [6], [9], and [11] or the complexity in the circuitry implemented in [7] lead to consider the current drive approach. More recently, a new technique allowing post pull-in operation of the moveable electrode, of an electrostatic microactuator, has been proposed by the authors [11]–[15]. The use of current pulses to inject the charge required to fix the electrode in post pull-in positions has been also tested showing good agreement with simulations. A further step is presented in this paper, a new sensing method, suitable for capacitive sensors, is conceptually introduced based on the use of current sources. The use of current sources applied to sense signals from a capacitive sensor has been reported in [16], [17]. However, with a totally different focus, the method presented here is addressed to control the charge injected in the capacitive sensor, and afterwards it senses the external excitation. To introduce the current approach, the paper has been organized in different sections comprising the transducer, the theoretical study, and the method proposed. First, the transducer structure is presented. The ac responses for the voltage and the charge drive mode are studied in Section III. The new method is presented and studied by simulations in Section IV. Finally, the main results are collected in Section V. 1530-437X/03$17.00 © 2003 IEEE