In the past, capacitive sensors for automotive electronics were rarely used, as they were deemed to be hard to control, difficult to read out, sensitive to aging, and temperature dependent. On the other hand, their favorable manufacturing costs, simple shape adaptation, and low power consumption are attractive attributes that provide motivation for their use. The emergence of a new measurement technique enables the number of capacitive sensors in a car to increase dramatically.
Macroscopically, capacitive sensors are usually analyzed by converting their capacitance into another physical variable, such as voltage, time, or frequency. Microscopically, capacitive sensors have existed in cars for a long time; micromechanical acceleration sensors are based on this principle. These are usually used to detect charge transfers.
A new method for sensing capacitance uses a modified input stage of a sigma-delta converter to detect an unknown capacitance and convert it into a digital value. This method, using capacitance-to-digital converters, (CDC), is explained herealong with several capacitive sensor principles that can be used in the car. Finally, an alternative method is outlined.
To visualize the CDC method, we must take an introductory look at the principle of sigma-delta conversion. Below is a simplified sigma-delta converter diagram.
In order to clearly understand its operation, we first look at the integrator's input, which must remain zero over long time intervals. Small short-term jumps will be converted into ramps. The zero average is achieved by raising the output of the reference branch to the same level as the input branch, which in turn is affected by the comparator's output. This switches the reference to the subsequent capacitance with a logical 1.
The capacitance is charged and applied to the integrator in reverse, so that the negative reference voltage is applied to the integrator. A high voltage at the input therefore causes a large number of logical ones, which in turn frequently apply the (negative) reference. The density of ones is converted by the following digital filter into a digital numerical value. A classic sigma-delta converter compares an unknown voltage with a known voltage and uses two known (usually equal) capacitances to do this.
In reality, charges are compared, Therefore, the capacitances can be compared using Q=C*V, if both voltages are known (equal voltages are used in this case). A synchronized voltage signal must also be applied to the input branch, as shown below in this capacitance-to-digital converter.
A number of advantages result from this method. Due to the close relationship with sigma-delta converters, their well-known properties can be modified and adopted. These include high noise suppression, high resolution at relatively low frequencies, and cost effective realization of high accuracy.
Sigma-delta converters, almost without exception, have a similar input structure, so various special structures can be adapted for specific measuring tasksfor example, particularly low current input, maximum accuracy, or higher cut-off frequency.
A further advantage becomes clear if we take a closer look at the above figure. Parasitic capacitances do not play any role in the initial approximation. A parasitic capacitance tending towards zero at node A has zero potential. Node B is not zero, but it is fed with a defined, low-resistance potential, so a parasitic capacitance at this node will charge to an average value that does not affect the measured result. A parasitic capacitance from node A to node B is always parallel to the measuring element, and will always appear as an offset.
Available capacitance-to-digital converters can provide very high performance. The AD7745 from Analog Devices, for example, achieves 24-bit resolution and 16-bit accuracy.
Previous capacitance analysis systems required comparatively high measurement capacitance and a large capacitance change when touched. This requirement for an adequately large change often caused problems for sensor manufacturers that would not occur with smaller capacitive sensors. Typical 150-pF moisture sensors, for example, are not only considerably more expensive (because they are larger), but are also more susceptible to errors and have lower long-term stability.
The capacitance of a capacitor can be calculated on the basis of its structure as
C = εoεr A/d
where εo is the dielectric constant of free space, εr is the dielectric constant of the material, A is the usable plate area, and d is the distance between the two electrodes. Other than a few exceptions, such as pressure sensors, all capacitive sensors use changes in the plate surface or the dielectric to measure changes in capacitance. Most sensors can be divided into two classes: those in which the plate area (geometry) changes (i.e. level sensors or displacement sensors), and those that rely on a change in εr(i.e. proximity or moisture sensors).