# Integrated Capacitive PGAs in ADCs, Part 2: Practical Examples

Editor’s note: I am delighted to bring you part 2 of a two-part tutorial on Integrated Capacitive PGAs in ADCs by our two guest bloggers Miguel Usach and Gerard Mora-Puchalt, from Analog Devices, Inc.

In Integrated Capacitive PGAs in ADCs, Part 1: Redefining Performance we discussed Capacitive PGA Circuit Description and Functionality.

Practical Examples

In a Wheatstone bridge, the common mode voltage is defined by the impedance connected in each of the legs, and is proportional to the applied power supply. Weigh scale applications implement this sensing topology due to the benefit of linear sensing in strain gauges, Figure 8 shows a half-bridge type II.

Figure 8 Weigh-scale with strain gauge in Wheatstone topology

The sensitivity of a strain gauge is typically 2 mV/V. The higher the Wheatstone supply, the higher the sensitivity obtained. To increase the dynamic range of the strain gauge and maximize the SNR, the bridge may be powered at higher supplies than the ADC.

In a resistive PGA, to maximize the dynamic range, the bridge should be powered at the same supply voltage as the ADC supply, while in a capacitive PGA the bridge can be powered at almost twice the ADC supply voltage.

For example, powering the ADC at 3.3 V, and assuming a buffer headroom of 100 mV, the improvement of a capacitive PGA over a resistive PGA is summarized in Table 1, assuming a maximum PGA gain of 128 (higher PGA gains are quite uncommon)

Table 1 Comparative of resistive and capacitive PGA in a Wheatstone bridge assuming standard supplies and gains

Another probable issue is a potential difference between grounds when the bridge is connected at some distance from the ADC. This could shift the common mode voltage, unbalance the ADC input common mode with respect to the bridge and reduce the maximum allowable gain in the resistive PGA.

A possible way to match the capacitive PGA performance with the resistive PGA is by powering the bridge at a higher supply voltage. For instance, powering the bridge with bipolar supply, +/-3.3 V, to increase the sensitivity to the strain gauge at the expenses of increase system complexity and power dissipation.

Another case example that could benefit from a capacitive PGA is a temperature measurement using resistance temperature detectors (RTD) or thermocouples.

A popular RTD resistor, such as the PT100, may be used to sense the temperature directly, or indirectly sensing the cold junction of a thermocouple, as shown in Figure 9.

Figure 9 Typical thermocouple setup

The PT100 is offered with different wires per element, being the most popular and cost effective 3-wire configuration.

A conventional way to measure the temperature while cancelling the lead error is proposed in Figure 10. In this example the internal current sources of the AD7124-8 Sigma-Delta ADC with PGA, drive two wires of the RTD with the same current, generating an offset error equal in both leads and proportional to the lead resistance.

Due to the small value of the lead resistance and the currents provided by the AD7174-8 to minimize the self-heating effect, the offset voltage generated in RL3 is close to the negative rail, significantly reducing the maximum allowable gain in a resistive PGA as its input common mode would also be very close to the rails as opposed to a capacitive PGA that will set internally the common mode voltage to half of the supply rails, allowing for a higher gain configuration and, therefore, increasing overall the dynamic range.

The proposed solution significantly reduces the complexity of the system and the hardware connections, as the third cable should not be returned to the ADC PCB and can be connected to ground near the RTD location.

Figure 10 3-Wire RTD measurement.

To increase the precision on the temperature measurement, 4-wire measurements is preferred. In this case, only one current reference is used. To avoid the imprecision on the current source and generate a ratiometric measurement with a precision resistance used as a reference, voltage reference of the ADC is generated by the precision resistance, as shown in Figure 11.

Figure 11 Ratiometric 4-wire RTD measurement

The value of the external precision resistor is chosen so that the maximum voltage generated across the RTD equals the reference voltage divided by the PGA gain.

In the case of a resistive PGA, the voltage generated on the precision resistance should be around 1.65 V, otherwise the common mode voltage will limit the maximum gain. The consequence is that maximum gained signal should be equal to 1.65 V.

In a capacitive PGA, the voltage generated in the precision resistance is independent of the maximum selectable gain of the amplifier. The VREF can be as high as rails, increasing the dynamic range and precision of the measurement.

Table 2 summarizes the maximum gain of a resistive PGA against a capacitive PGA, with a maximum current source of 500 uA to limit the Pt100 self-heating, assuming a Class B RTD, at a maximum temperature of 600o C, a maximum VREF of 2.5 V

Table 2 Comparative of resistive and capacitive PGA in a 4-wire RTD radiometric measurement

Conclusion

The capacitive PGA offers an important number of advantages compared with the resistive PGA. Critical specifications such as noise, common mode rejection, offset, gain error and temperature drifts are improved due to the inherent temperature stability and matching properties of the capacitors as gain elements.

Another key feature is the decoupling of the input common mode voltage from the amplifier internal common mode voltage. This is critical when the input signal to be amplified sits on a common mode voltage close to the rails. The resistive PGA selected gain would be severely restricted by its common mode limitation, or it would require higher supply rails or external components to re-bias the input signal to half of the rails. On the contrary, the capacitive PGA could handle this sensing scenario easily.