Lead Off Detection in Biopotential Signal Chains
In the last blog post, I discussed electrode models and the challenges of managing dry electrodes. Building off this information, let’s shift to the topic of electrode or lead (pair of electrodes) off detection and how this feature can impact the power and performance of the signal chain. Lead off detection is a method for providing some indication (typically a logic signal or interrupt) that a user is sufficiently connected to the electrodes such that a biopotential signal measurement can be made.
So how does this affect the power? For one, when a device such as a watch or chest strap is not being used, you wouldn’t want to burn power unnecessarily. Some indication of a “leads off” condition could enable the power down of components in the signal chain while in this state. Similarly a “leads on” condition could act as a wake-up signal to alert the microprocessor to power up components and/or start taking measurements again. See an example of this type of sequence using the lead off detection (LOD) and shutdown (SDN) pins of the AD8233 in Figure 1. The lead off detection circuitry also adds to the total power consumption. This may be more important if the function is still active during power down.
Figure 1 – Electrode connection and system wakeup sequence for the AD8233
Implementation of lead off detection will depend on the use of two or three electrodes and can also provide input biasing (see fifth blog). From a performance perspective, any circuitry touching the electrodes can affect the input impedance and input currents can lead to electrode polarization and an increase in differential offsets.
Two Electrode DC Lead Off Detection
Figure 2 shows a DC lead off detection implementation for a two-electrode solution. A pull-up resistor (RPU) is tied to supply voltage on one input and an equal pull-down resistor (RPD) is tied to ground on the other input. Rtotal represents the body impedance and the contact impedances for the two electrodes. A simplified way of looking at this, is if Rtotal is an open, the inputs to the instrumentation amplifier will pull apart and rail the output indicating “lead off”. This would happen if either electrode were disconnected. When both electrodes are connected, and Rtotal is approximated as a short, then the inputs will be biased to mid supply (+Vs/2).
Figure 2 – DC lead off detection and biasing for a two-electrode solution
In actuality, the electrode contact impedances will contribute to some differential offset and RPU and RPD should be made sufficiently large (say 10MΩ to start) to minimize current through the electrodes. The fact that the current is flowing through electrodes in equal and opposite directions means their polarization will be different. This is one reason I prefer the three-electrode solution discussed in the next section. The “lead off” indicator could also come from low power comparators. The comparators are used to monitor the voltages at the instrumentation amplifier inputs and allow for preset threshold voltages.
Three Electrode DC Lead Off Detection
In the three-electrode solution shown in Figure 3, the input bias (Vbias) is provided from a third electrode. Now each input has a pull-up resistor to the high supply such that if there is a loss of contact on electrodes 1 to 3 (Rc1-Rc3), separate comparators can detect which input electrode is compromised. By having the current flow in the same direction in Rc1 and Rc2, the polarization of the electrodes should track better and lead to a common mode shift rather than a large differential offset.
Figure 3 – DC lead off detection and biasing for a three-electrode solution
DC current sources could be used instead of pull-up resistors as well. This helps to take out the current dependency on the supply voltage and Vbias while also decoupling the input impedance from the current amplitude. The programmability of these currents as well as the comparator threshold voltage adds flexibility for different electrode types. The ability to disconnect the current sources enables a spot check vs. an always on approach.
AC Lead Off Detection
The use of AC Lead Off Detection can take advantage of the fact that the electrode model includes a capacitor in parallel with a resistor. This means the electrode impedance is frequency dependent and can be smaller at higher frequencies. AC current sources as shown in the ADPD6000 block diagram in Figure 4 allow for zero DC current (no electrode polarization) and can be set at a frequency outside the bipotential signal bandwidth (4kHz in this case). A source and sink current are switched back and forth between the two ECG channel inputs to generate the AC current. The amplitude of the corresponding voltage signal can be measured by the same ECG channel and synchronously demodulated for comparison against a programmable threshold voltage.
Figure 4 – AC lead off detection in ADPD6000
This is the last in the series of biopotential signal chain focused blogs, however we hope to bring more Precision Low Power topics in the future. In the meantime, continue to visit our Signal Chain page as well as our dedicated EngineerZone page for the latest content and collateral.