The circuit shown in Figure 1 is a fully isolated 12-bit, 300 kSPS data acquisition system using only three active devices.
The system operates from a single 3.3 V supply and can handle ±10 V input signals. After room temperature calibration, the total error does not exceed ±0.1% FSR over a temperature variation of ±10°C, making it ideal for a variety of industrial measurement applications.
The circuit's small size makes the combination an industry-leading solution for data acquisition systems where accuracy, speed, cost and size are critical. Data and power are isolated from each other, so the circuit has excellent high-voltage tolerance while also effectively avoiding ground loop interference common in harsh industrial environments.
The circuit consists of an input signal conditioning stage, an ADC stage, and an output isolation stage. The ±10 V input signal is level shifted and attenuated by the U1A op amp, which is one half of the dual AD8606 . The output of this op amp is 0.1 V to 2.4 V, matching the input range of the ADC (0 V to 2.5 V), with a margin of 100 mV to maintain linearity. A buffered reference voltage (VREF =2.5 V) from the ADC is used to generate the required offset. The resistor values can be modified to accommodate other common input ranges as described later in this circuit note.
The circuit design supports single power supply. The AD8606 is specified for a minimum output voltage of 50 mV (2.7 V supply) and 290 mV (5 V supply), a load current of 10 mA, and a temperature range of −40°C to +125°C. With a 3.3 V supply, load current less than 1 mA, and a narrower temperature range, a conservative estimate of the minimum output voltage is 45 mV to 60 mV.
Taking into account device tolerances, the minimum output voltage (lower end of range) is set to 100 mV to provide a safety margin. The upper limit of the output range is set to 2.4 V to provide 100 mV of headroom for positive swings at the ADC input. Therefore, the nominal output voltage range of the input op amp is 0.1 V to 2.4 V.
The other half of the AD8606 (U1B) is used to buffer the internal 2.5 V reference voltage of the AD7091R (U3) ADC.
The AD8606 was chosen for this application because of its low offset voltage (65 μV maximum), low bias current (1 pA maximum), and low noise (12 nV/√Hz maximum) characteristics. At 3.3V supply, power consumption is only 9.2 mW.
The output stage of the op amp is followed by a single-pole RC filter (R3/C9) to reduce out-of-band noise. The cutoff frequency of the RC filter is set to 664 kHz. An optional second-order filter (R4, C10 and R1, R2, C11) can be added to further reduce the filter cutoff frequency in the presence of low-frequency industrial noise. In such cases, the sampling rate of the AD7091R can be reduced due to the smaller signal bandwidth.
The AD7091R 12-bit 1 MSPS SAR ADC was chosen because of its ultra-low power consumption of only 349 A at 3.3 V (1.2 mW), which is significantly lower than any competing ADC currently on the market. The AD7091R also has a built-in 2.5 V reference with a typical drift of ±4.5 ppm/oC. The input bandwidth is 7.5 MHz, and the high-speed serial interface is SPI-compatible. The AD7091R is available in a small 10-pin MSOP package.
When powered from a 3.3 V supply, the total power dissipation of this circuit (excluding the ADuM5401 isolator) is approximately 10.4 mW.
Galvanic isolation is provided by the ADuM5401 quad-channel digital isolator (Grade C). In addition to isolating the output data, the ADuM5401 provides an isolated 3.3 V power supply for this circuit. The ADuM5401 is not required for normal operation of the circuit unless isolation is required. The ADuM5401 quad-channel 2.5 kV isolator integrates a DC/DC converter and is available in a small 16-pin SOIC package. The ADuM5401 consumes approximately 140mW at a 7 MHz clock frequency.
The AD7091R requires a 50 MHz serial clock (SCLK) to achieve a 1 MSPS sampling rate. However, the ADuM5401 (Grade C) isolator has a maximum data rate of 25 Mbps, which corresponds to a maximum serial clock frequency of 12.5 MHz. In addition, the SPI port requires that the trailing edge of SCLK drives the output data to the processor, so the total bidirectional propagation delay of the ADuM5401 (120 ns maximum) limits the upper clock limit to 1/120 ns = 8.3 MHz.
Although the AD7091R is a 12-bit ADC, the serial data is also formatted as 16-bit words to be compatible with the processor serial port requirements. Therefore, the sampling period TS includes the conversion time of the AD7091R650 ns plus 58 ns (additional time required by the data sheet, t1 delay + tQUIET delay), plus 16 clock cycles for the SPI interface data transfer.
T S = 650 ns + 58 ns + 16 × 120 ns = 2628 ns
f S = 1/T S = 1/2628 ns = 380 kSPS
To provide a safety margin, it is recommended to set the SCLK and sample rate to a maximum value of 7 MHz and 300 kSPS respectively. The digital SPI interface can be connected to the microprocessor evaluation board using a 12-pin Pmod compatible connector (Digilent Pmod specification).
circuit design
The circuit shown in Figure 2 attenuates and level-converts a −10 V to +10 V input signal to an ADC input range of 0.1 V to 2.4 V.
The transfer function is found using the superposition principle:
Calculation of Gain, Output Offset, and Resistor Values
If the input voltage range is ±10 V, the calculation is as follows.
In the actual circuit, resistor values closest to existing standards were chosen for resistors R4 and R5. The selected values are R4 = 52.3 k¬ and R5 = 12 k¬. Note that R1 = R4, R2 = R5.
If these values are chosen carefully, the total error caused by using alternative standard value resistors can be reduced to less than a few percent. However, the output of the U1A op amp at a ±10 V input should be recalculated using Equation 1 to ensure that the required margin is maintained.
The absolute accuracy of such circuits depends primarily on the resistors, so gain and offset calibration is required to eliminate errors caused by substitute standard value resistors and resistor tolerances.
Calculate resistance for different input ranges
For input ranges other than ±10 V, complete the following calculation steps.
Replace R and R0 in Equation 17 and Equation 18 with the values defined in Equation 2 and solve both equations to find the R4/R6 ratio.
Choose the value of resistor R6. Calculate R4 from the R4/R6 ratio. Obtain the R4 and R6 values and calculate R5 using Equation 2 and the R4/R6 ratio. Calculate R2 and R1 by Equation 16. R1 = R4 can be appropriately chosen and R2 calculated.
Effect of Resistor Temperature Coefficient on Total Error
Equation 1 shows that the output voltage is related to the following five resistors: R1, R2, R4, R5, and R6. The full-scale output voltage at TP1 is sensitive to small changes in the resistance of each of these five resistors, and its sensitivity is calculated through a simulation program. The input voltage to the circuit is +10 V. The calculated sensitivities are S R1 = 0.19, S R2 = 0.19, S R4 = 0.39, S R5 = 0.11, S R6 = 0.50. Assuming that the temperature coefficients combine in a root-sum-square (rss) fashion, the total full-scale drift using a 100 ppm/°C resistor is approximately:
Full scale drift =
=100 ppm/°C √(S R1 2 + S R2 2 + S R4 2 + S R5 2 + S R6 2 )
=100 ppm/°C √(0.19 2 + 0.19 2 + 0.39 2 + 0.11 2 + 0.50 2 )
= 69 ppm/°C
A full-scale drift of 69 ppm/°C corresponds to 0.0069% FSR/°C. Using a 25 ppm/°C resistor reduces the drift error to 0.25 × 69 ppm/°C = 17 ppm/°C, or 0.0017% FSR/°C.
Effect of active component temperature coefficient on total error
The DC offset of the AD8606 op amp and AD7091RADC is removed by the calibration procedure.
The offset drift of the AD7091R ADC's built-in reference is 4.5 ppm/°C typical and 25 ppm/°C maximum.
The offset drift of the AD8606 op amp is 1 V/°C typical and 4.5 V/°C maximum.
U1A The error due to the AD8606 input is referenced to the 2.3 V output range and is therefore 2 ppm/°C. The error due to the U1B reference voltage buffer is also approximately 2 ppm/°C based on 2.5 V.
The total drift error result is summarized in Table 1. These errors do not include the ±1 LSB integral nonlinearity error of the AD7091R.
Note that if 50 ppm/°C or 100 ppm/°C resistors are used, the largest source of total drift is resistor drift, and drift from active components is negligible.
Table 1. Error due to temperature drift
source of error | total error |
Resistance (1%, 100 ppm/°C) |
±0.0069% FSR/°C |
AD7091R (ΔVVREF/ΔT = 25 ppm/°C) |
±0.0025% FSR/°C |
AD8606, U1A (ΔVOS/ΔT= 4.5 μV/°C), 2 ppm/°C, Referenced to 2.3 V |
±0.0002% FSR/°C |
AD8606, U1B (ΔVOS/ΔT= 4.5 μV/°C), 2 ppm/°C, Referenced to 2.5 V |
±0.0002% FSR/°C |
Total FSR Error Temperature Coefficient (100 ppm/°C resistor) |
±0.0098% FSR/°C |
Total % FSR Error for ΔT=±10°C (100 ppm/°C resistance) |
±0.098% FSR |
Total % FSR Error for ΔT=±10°C (25 ppm/°C resistance) |
±0.046% FSR |
Test data before and after two-point calibration
To perform a two-point calibration, first apply −10 V to the input and record the ADC output code as Code_1. Then, apply a current of +10 V to the input terminal, and record the ADC output code as Code_2. The gain coefficient is calculated by the following formula
The pre-calibration error is obtained by comparing the ideal transfer function calculated using the nominal values of the components with the uncalibrated actual circuit transfer function. The tolerance of the resistors used in the actual measured circuit is ±1%. Test results do not include temperature changes.
Figure 3 shows the percent error (FSR) test results before and after calibration at ambient temperature. As shown in the figure, the maximum error before calibration is approximately 0.23% FSR. After calibration, the error is reduced to ±0.03% FSR, which is roughly equivalent to the 1 LSB error of the ADC.
PCB layout considerations
For a complete documentation package, including schematics, board layout and bill of materials (BOM), please refer to www.analog.com/CN0335-DesignSupport .
High voltage capability
This PCB is designed according to the 2500 V basic insulation specification. High voltage testing above 2500 V is not recommended. Caution must be exercised when using this evaluation board at high voltages, and this PCB must not be relied upon for safety functions as it has not been high potential tested (also known as high voltage testing or withstand voltage insulation testing) and has not been safety certified.
Blockdiagram
All reference designs on this site are sourced from major semiconductor manufacturers or collected online for learning and research. The copyright belongs to the semiconductor manufacturer or the original author. If you believe that the reference design of this site infringes upon your relevant rights and interests, please send us a rights notice. As a neutral platform service provider, we will take measures to delete the relevant content in accordance with relevant laws after receiving the relevant notice from the rights holder. Please send relevant notifications to email: bbs_service@eeworld.com.cn.
It is your responsibility to test the circuit yourself and determine its suitability for you. EEWorld will not be liable for direct, indirect, special, incidental, consequential or punitive damages arising from any cause or anything connected to any reference design used.
Supported by EEWorld Datasheet