Circuit provides linear resistance-to-time conversion

Resistance-based transducers, such as strain gauges and piezoresistive devices, find common use in the measurement of several physical parameters. For applications in which digital processors or microcontrollers serve for data acquisition and signal processing, the transducer's response must assume a form suitable for conversion to digital format. It is desirable to convert the resistance change of such sensors into a proportional frequency or a time interval so that you can easily obtain an output in digital form, using a counter/timer. The circuit of Figure 1 linearly converts the sensor resistance, RS, into a proportional time period. The circuit is essentially a relaxation oscillator, comprising a current source, a bridge amplifier, a comparator, and a flip-flop. The current, IS, divides in the paths of R1 and R2 as if the two resistors were connected in parallel. Assuming ideal op amps, the circuit functions as an oscillator when RX (R4+RS) is greater than R1R3/R2.

The circuit produces waveforms at the input and output of the comparator, IC2 (Figure 2).

T1 and T2 are the time intervals for which the comparator's output assumes levels VS1 and ÐVS2, respectively. The output voltage from IC2, with its levels changed via a zener-diode circuit, serves as clock input to a D flip-flop. From the 7474 flip-flop, you obtain a square-wave output that is high and low alternately for a time period T=4C(R2RXÐR1R3)/R1. This equation indicates that the circuit converts a change in sensor resistance into a proportional time period ÆT with sensitivity ÆT/ÆRS=4C(R2/R1). The following salient features of Figure 1 merit mention:

• The sensor is grounded; you can easily vary the conversion sensitivity by varying either R1 or R2.


• You can adjust the offset value, T0 (about which changes in T occur because of a change in the sensor's resistance), by changing either R3 or R4 without affecting the conversion sensitivity.


• The offset voltages of the op amps alter T1 and T2 in opposite ways, such that their overall effect on T(T1+T2) is not appreciable.


• Thanks to the current source, the output is largely insensitive to noise voltages in the line of the current source and to changes that occur in VS1 and VS2.


Consider the example of converting a Pt-100 (platinum RTD) sensor in the range of 119.4 to 138.51½, which corresponds to a temperature range of 50 to 100¡C, into time periods of 10 to 12.5msec. The design is simple. Because the current through the sensor is a fraction of IS, IS should be low enough to keep the self-heating error to an acceptably low level. This design uses an IN5287 current regulator; it provides an IS of approximately 0.33mA and has a dynamic impedance better than 1.35M½. For a better current source, you could use a circuit based on a voltage-regulator IC. In the next step, with suitable and practical fixed values for R1 and C, you adjust R2 until you obtain the needed sensitivity: 130.82µsec/½. Following this step, with a fixed R4, you adjust R3 to obtain the offset required in the output (T). Figure 1 shows the values of components for this example. The resistors all have 1% tolerance and 0.25W rating, and C is a polycarbonate capacitor.