In a hospital environment, many types of diagnostic and clinical equipment are used to monitor and control patients’ health. Due to the fact that increasing numbers of electronic devices are transmitting data, the electromagnetic interference (EMI) between the different pieces of equipment can cause malfunctioning of the medical devices, and this is becoming a serious concern.

In a medical environment, EMI can be caused by a visitor’s mobile phone for example, or by the radio transmitters used by an ambulance. Other interference effects can be due to patients’ electric razors or the call systems for the nurses. Furthermore, devices used by the medical team such as cell phones, walkie-talkies, personal digital assistants, and two way pagers can be a source of emissions. Finally, the radiating medical instruments such as X-ray machines, CAT (computer axial tomography) scanners, and MRI (magnetic resonance imaging) machines are also potential sources of EMI.

Effects of EMI
EMI can cause medical equipment to fail with more or less severe consequences. It can cause loss of control or create errors in a patient’s data readings such as pulse rate or gas concentration. Even more dangerously, it can produce errors in the amount of medicine administered to a patient by an infusion system.

To avoid such concerns, the medical equipment is tested for susceptibility and has to comply with various standards. Common methods of preventing the harmful effects of EMI are cable shielding and filtering. In addition, regulations are enforced in medical environments to reduce the risk of EMI problems like checking to determine whether the equipment is compliant to standards or restricting the use of mobile phones or other wireless equipment.

To further protect the patient, EMI hardened components should be used to design the signal path. This will be illustrated using the example of an electrocardiogram (ECG) signal conditioning system. An ECG is a recording of the heart’s electrical activity over time. This signal is measured by connecting three small electrodes to the human body. The ECG signal presented in Figure 1 is characterized by five specific points, P, Q, R, S, and T, which permit the diagnosis of possible heart diseases.

1. ECG signal. The three electrodes are placed on the body surface to acquire the ECG signal.

The signal collected through the electrodes can range from a 400 µV to 5 mV peak, with 3 dB corner frequencies at 0.05 and 100 Hz. Due to its very low amplitude, this signal is generally disrupted by various types of interference such as electrode contact noise, power-line noise (50 Hz), respiration, muscle activity, and as previously mentioned, interference from other electronic devices. This implies that the signal path must deal with different noise sources. To reject dc noise, high-pass filters can be implemented. The main concern is the 50-Hz noise which is in the same frequency range as the signal of interest to the doctors. Use of an instrumentation amplifier has proven highly effective in eliminating common-mode noise. Indeed, this configuration is ideal because it rejects common-mode voltage while amplifying differential voltages, which allows the separation of the low-level signal from the background noise.

2. ECG signal conditioning. This instrumentation amplifier was designed using the LMP2021 precision op amp.

As illustrated in figure 2, the instrumentation amplifier is implemented using the LMP2021 precision operational amplifier that offers a low input offset voltage (0.4 μV typical) and near zero input offset voltage drift (0.004 μV/°C), low input voltage noise, and high open loop gain. The proprietary continuous correction circuitry enables quality CMRR and PSRR performance and removes the flicker noise component. This auto correction eliminates the need for calibration in many circuits. The op amp’s 260 nVpp (0.1 to 10 Hz) of input voltage noise and no 1/f noise component make it suitable for low-frequency applications such as non-invasive low-frequency medical instrumentation applications.

In the signal path, the part that’s the most sensitive to EMI is the interface between the sensor and the op amp. In this section, the signal is analog, with low amplitudes and the wires are long, resulting in an increased susceptibility to interference. The interface between the op amp and the analog-to-digital converter (ADC) is less sensitive, as the signal has higher amplitude levels because the amplification and the wires are generally shorter.

As the critical point is the interface between the sensor and the op amp, analog semiconductor vendors have introduced product solutions with integrated EMI filters that maintain the accuracy of analog systems by reducing the effects of radio frequency (RF) interference.

To allow for the characterization of EMI hardened op amps, a new parameter in the datasheet is needed that provides a quantitative description of the EMI performance of op amps. This quantitative measure shows how well the op amp rejects the EMI, and enables the comparison and also ranking of op amps based on their EMI robustness and performance. Much like CMRR, EMI rejection ratio (EMIRR) is specified as the ratio of the change in an applied RF signal to the resulting change in the offset voltage.

The EMIRR is defined as:

where, Vrf_peak is the amplitude of the applied un-modulated signal, and ΔVos is the resulting input referred offset voltage shift.

3. ECG signal conditioning. Shown is a block diagram of the ECG system.

Figure 3 shows the block diagram of the complete signal path. Filters are then implemented to suppress the unwanted signals. The analog high-and low-pass active filters can be implemented using the Sallen-Key topology, built around op amps with resistors and capacitors.

The analog signal is then converted to digital with the ADC161S626, a 16-bit successive-approximation register ADC. The part has a minimum signal span accuracy of ±0.003% over the temperate range of −40°C to +85°C. The converter features a differential analog input with a common-mode signal rejection ratio of 85 dB.

When building the complete system, isolation is needed for the patients’ security. This can be done with either galvanic isolation, photo-opto coupling, capacitive coupling, or magnetic coupling.

Carine Alberti is a precision systems marketing engineer for healthcare and sensing and motor control at National Semiconductor’s European Headquarters in Fürstenfeldbruck, near Munich/Germany. Carine joined the company in 2005 as an application engineer for amplifiers. She received her electrical engineering degree from ENSERG (École Nationale Supérieure d'Électronique et de Radioélectricité de Grenoble) in France.