SBAA532A February 2022 – March 2024 ADS1119 , ADS1120 , ADS1120-Q1 , ADS112C04 , ADS112U04 , ADS1130 , ADS1131 , ADS114S06 , ADS114S06B , ADS114S08 , ADS114S08B , ADS1158 , ADS1219 , ADS1220 , ADS122C04 , ADS122U04 , ADS1230 , ADS1231 , ADS1232 , ADS1234 , ADS1235 , ADS1235-Q1 , ADS124S06 , ADS124S08 , ADS1250 , ADS1251 , ADS1252 , ADS1253 , ADS1254 , ADS1255 , ADS1256 , ADS1257 , ADS1258 , ADS1258-EP , ADS1259 , ADS1259-Q1 , ADS125H01 , ADS125H02 , ADS1260 , ADS1260-Q1 , ADS1261 , ADS1261-Q1 , ADS1262 , ADS1263 , ADS127L01 , ADS130E08 , ADS131A02 , ADS131A04 , ADS131E04 , ADS131E06 , ADS131E08 , ADS131E08S , ADS131M02 , ADS131M03 , ADS131M04 , ADS131M06 , ADS131M08
Figure 3-1 illustrates how bridge measurements are typically made with a ratiometric reference configuration. The output of the bridge is measured by the ADC while a single source is used as both the bridge excitation voltage and the ADC reference voltage.
The ADC samples the input voltage, VIN, and compares it against the reference voltage, VREF. VIN is the voltage difference between VSIGNAL+ and VSIGNAL– (or AINP and AINN) while VREF is the difference between the voltage at VSENSE+ and VSENSE– (or REFP and REFN). The ADC generates an output code proportional to VREF as per Equation 14:
Equation 14 substitutes AVDD for the VREF term because AVDD is connected to REFP and REFN is grounded in Figure 3-1. Also, recall that Equation 13 states that VOUT is equal to VEXCITATION multiplied by ΔR divided by R. In Figure 3-1, AVDD = VEXCITATION+ - VEXCITATION- while VOUT (Bridge) = VIN (ADC). These substitutions yield Equation 15:
Substituting Equation 15 for the VIN term in Equation 14 results in Equation 16:
The output code in Equation 16 is directly proportional to ΔR. Moreover, Equation 16 shows that the exact values of AVDD and VIN are unnecessary. Instead, the output code is directly proportional to the strain on the bridge.
One of the benefits of a ratiometric measurement is that the measurement is relatively invariant to changes in VREF. This is also shown in Equation 16, where the output is proportional to ΔR / R and is therefore independent of the exact value of VREF or VEXCITATION. Therefore, the ratiometric measurement is less susceptible to any VEXCITATION drift over time and temperature. Any noise from the VEXCITATION source should also cancel out assuming that the noise at the reference input and the measurement input are correlated. Typically, these noise sources correlate well if the filter bandwidth for the reference input and the measurement input are the same.