Brain-burning analysis of a different ECG front end, IoT, WSN... can be easily mastered

However, there is an ECG front-end IC that can operate at 50 uA supply current and has a small 2 mm × 1.7 mm WLCSP package that is worth considering when designing IoT node applications.

If you dig a little deeper, you’ll find that the flexible architecture of this ECG front-end IC is essentially an instrumentation amplifier (IA) and a few op amps that can be configured to form some useful ultra-low-power signal processing circuits that are useful in more than just medical or healthcare applications.

A simplified single-lead electrocardiogram (ECG) front end is shown in Figure 1.

It consists of an indirect current mode IA with independent transfer function:

In this front-end example, the fixed gain is 100. The reference source of the IA is driven by a high-pass amplifier (HPA) configured as an integrator in a feedback network with its input connected to IAOUT and the cutoff frequency set by external capacitors and resistors. The HPA will force HPDRIVE to whatever voltage is required to keep HPSENSE and IAOUT at the reference voltage. This circuit forms a first-order high-pass filter:

For diagnostic grade ECGs, the cutoff frequency is typically set to 0.05 Hz, while for healthcare applications that only detect heart rate, 7 Hz may be appropriate. The high-pass filter function solves the problem of how to amplify the high frequency ECG signal (1 mV to 2 mV) while rejecting the large DC half-cell potential (caused by the electrode/skin contact) and the low frequency baseline drift associated with ECG measurements. Since the DC half-cell potential (up to 300 mV) rejection occurs at the input of the IA, this architecture can achieve a large gain. Another benefit is that the offset and offset drift of the IA can be suppressed. Monitoring the HPDRIVE with respect to the reference voltage will show an inverted version of the input offset that is automatically corrected.

Although this design was originally intended for ECG applications, virtually any application that needs to amplify low frequency small signals (IA bandwidth < 1 kHz) can benefit from its low power consumption and small size. If dc measurements are to be made, only a simple modification of this circuit is required.

Figure 2 shows a DC-coupled IA with a fixed gain of 100. This is done by removing the R and C from Figure 1 and shorting HPSENSE to HPDRIVEA, making the HPA a unity-gain buffer. This also forces the IA reference to maintain the reference voltage. The offset voltage of the IA should be considered in this case.

If a gain of 100 is too high, or a bandwidth of 1 kHz is too low, the circuit can be modified as shown in Figure 3.

The HPA is now configured as an inverting amplifier with a gain of –R2/R1, and its input is the feedback from IAOUT. The new transfer function can be simplified as follows:

Gains less than 100 can be achieved by configuring the HPA as an attenuator (R2 < R1). Since the differential input is limited to 300 mV, it is recommended that the gain should not be less than 10 to ensure the stability of the circuit. The following table lists some gain configurations that can be considered.

Table 1. DC-coupled IAs with different gain and bandwidth configurations.

If dc accuracy is still important, leave the IA gain at 100 and modify the circuit as shown in Figure 4 to compensate for the offset of the IA and any attached sensors.

The adjusted transfer function is as follows:

VTUNE is the source voltage used to correct the offset voltage, which can be provided by a PWM filtered signal from a microcontroller or directly driven by a low-power DAC. The HPA is still configured as an inverting amplifier with a gain of –R2/R1, which can be used to further adjust the offset correction range and resolution. Decomposing VIN and then substituting it into the above formula, the target transfer function can be obtained:

The overall offset can be compensated by adding sensors that do not have VSIGNAL applied. Simply measure IAOUT relative to the reference and adjust (R2/R1) VTUNE until the voltage is close enough to zero.

Before using the above circuit configuration for low power IoT design, it is also important to understand the other parts of the AD8233 ECG front-end solution. The detailed circuit is shown in Figure 5.

The digital input FR supports a fast restore function, which is very beneficial for the AC-coupled circuit in Figure 1. During startup or when a DC step event occurs at the input, the external capacitor needs some time to charge. In this case, the IA will enter rail-to-rail mode until the integrator has settled. Automatic fast restore detects this event and then switches to a smaller resistor in parallel with the external resistor for a certain period of time, greatly speeding up the settling process. The SW pin is used to quickly settle a second external high-pass filter if necessary.

The AC/DC digital input determines the lead-off detection method used in ECG applications, but can also be used for wire break detection where the input is other sensors. If properly configured, the digital output LOD will indicate when one of the IA inputs is disconnected from the sensor.

In addition to its small size and low active power consumption, the AD8233 also features a shutdown pin (SDN) that reduces the total supply current to less than 1 uA. This is very convenient for applications where sensor measurements are not often performed and can significantly extend the overall battery life. Even in shutdown mode, the wire break detection will remain functional.

Now that we have a more detailed understanding of the overall AD8233 chip, let’s look at a few different ideas for sensor applications. Table 2 lists a guide to getting started with building non-ECG circuits.

Table 2: Getting Started with the AD8233 for Non-ECG Applications

A good example of a Wheatstone bridge-based pressure sensor application that would be suitable for a fixed gain of 100 and the offset correction circuit of Figure 4 is a Wheatstone bridge-based pressure sensor application. This bridge naturally sets the input common-mode voltage to +Vs/2. The bridge can be driven by REFOUT or an uncommitted op amp (depending on the measurement range and required current) so that the supply current to the bridge is disabled in shutdown mode. Figure 7 shows an example circuit.

Another application that can benefit from the circuit of Figure 4 is temperature measurement using thermocouples. Type K thermocouples are nearly linear over a wide temperature range, with a Seebeck coefficient of approximately 41 uV/°C at room temperature (25°C). Assuming the reference or cold junction is compensated, the IA output will be a gain-over signal of the measurement junction, ~4.1 mV/°C (a NIST lookup table can be used for more accurate results). The output of a thermocouple is simply the difference between the measurement and reference junctions, so an equal reference junction drift must be added to cancel it out.

To begin this process, determine the desired reference junction temperature range and the expected drift from the NIST table. For example:

If an accurate temperature sensor is placed at the reference, the measurement can be fed back to VTUNE and adjusted by –R2/R1 to obtain the appropriate drift. Note that the temperature sensor should be drifted negatively, or the IA inputs should be swapped to ensure a positive drift at the IA output. To isolate the offset and drift correction, the circuit can be broken down into a summing node where the offset at VTUNE2 is fixed by –R2/R3. The updated transfer function is as follows:

The modified circuit is shown in Figure 8. Note that the input common-mode voltage is set to +Vs/2 by the 10 MΩ pull-up resistor on +IN and the 10 MΩ pull-down resistor on –IN. This configuration enables the leads-off detection feature of the AD8233 by pulling +IN up to +Vs in the event of a wire-off event. This condition can be monitored via the LOD pin. The AD8233 also has an integrated RFI filter to help with any high frequency pickup from the thermocouple. Adding additional resistors in series with the input can reduce the cutoff frequency.

An in-depth analysis of the AD8233 shows that its applications extend beyond the ECG front end. The device's unmatched combination of effective low power consumption (50 uA), a small 2 mm x 1.7 mm WLCSP package, shutdown pins, and a flexible architecture enable designs that are smaller, lighter, and have longer battery life. So, the next time you're working on an IoT, WSN (wireless sensor network), or any other low-power design, consider the AD8233 device and think about what circuits you can implement with it. Your battery life may depend on it.