Abstract: This paper introduces the design method of a digital sensor for surface electromyography (EMG) signals. Based on the characteristics of EMG signal generation and the basic requirements of acquisition technology, the shape and correct placement of the electrodes are studied, and a preamplifier circuit is designed using the INA128 instrumentation amplifier. An active filter is designed, and a serial A/D converter chip is used to output the digital signal. Experiments show that this method can improve the signal-to-noise ratio, reduce noise, and effectively extract surface EMG signals.
Keywords: surface electromyography signal; digital sensor ; preamplifier circuit
Abstract: Design of surface electromyography digital transducer was discussed. According to the special character of surface electromyography and basic need of collection technology, it had been explored to investigate the electrode shape and the right placing method. The instrumentation amplifier INA128 was adopted to design preamplifier circuit. Active filter was designed and serial A/D switch circuit controller was used to output digital signal. Experiments showed that the design can improve of signal and noise, reduce noise ratio and detect surface electromyography effectively.
Keywords: Surface electromyography, Digital transducer, Data acquisition
1 Introduction
Surface electromyography (sEMG) signals are one-dimensional voltage time-series signals obtained by guiding, amplifying, displaying and recording the bioelectrical changes of the neuromuscular system during voluntary and involuntary activities through surface electrodes. The amplitude is about 0-5000μV and the frequency is 0-1000Hz. The signal morphology has strong randomness and instability. Compared with traditional needle electromyography, sEMG has relatively low spatial resolution, but a larger detection space and better repeatability. It has important academic value and application significance for sports science research, rehabilitation medicine clinical and basic research [1].
The human body is a conductor, and power frequency interference, as well as external electric and magnetic fields, can generate measurement noise within the body, interfering with sEMG detection. Therefore, signal filtering and circuit shielding are key aspects of the design of digital surface electromyography (sEMG) sensors. The design consists of several parts: electrodes, amplifier circuit, filter circuit, and A/D converter.
2. Electrode Design
The electrode substrate in this paper is made of copper and silver-plated. It adopts a common bipolar design, with a reference electrode, also known as an unrelated electrode, inserted between the two electrodes to reduce noise and improve common-mode signal suppression. Differential amplification is used to eliminate noise from the power supply lines.
Electromyography (EMG) signals are detected by two electrodes. The two input signals are "subtracted" to remove the same "common-mode" components and only amplify the different "differential-mode" components. Any noise that is far away from the detection point will appear as a "common-mode" signal at the detection point; while the signal near the detection surface will appear different and will be amplified. Therefore, electric field noise that is relatively far away will be eliminated, while EMG signals that are relatively close will be amplified. Its accuracy is measured by the common-mode rejection ratio (CMRR) [2].
The transmission of electromyographic (EMG) signals within human tissue (a volume conductor) attenuates rapidly with increasing distance. Therefore, electrodes should be placed on the muscle belly where EMG signals are strongest to reduce EMG interference (crosstalk) from nearby muscles. Using smaller electrodes can improve selectivity, but it will increase the contact impedance between the electrode and the skin.
3. Design of Amplifier Circuit
Human muscle tissue is the signal source of electromyography (EMG) on the skin surface. The EMG it emits is transmitted to the skin surface through the body resistance of the subcutaneous soft tissue. The body resistance is about several hundred ohms. However, the contact impedance between the surface electrode and the skin is relatively high, about several thousand ohms. The contact resistance is also affected by a variety of factors such as the tightness of the contact, the cleanliness of the skin, humidity, and seasonal changes, and varies greatly [3]. It can be seen that for the amplifier, the EMG signal source is a high internal resistance signal source.
When designing the electromyography (EMG) signal amplification circuit, the following issues were emphasized: 1. High gain: The amplitude of surface EMG signals is approximately in the range of μV to mV, which is an extremely weak signal. It needs to be amplified to about one volt for practical use, so the amplifier gain was set to 80dB. 2. High common-mode rejection ratio (CMRR): Surface EMG signal acquisition is susceptible to interference from 50Hz power supply and other high-frequency electrical noise. However, these interference signals appear as in-phase signals at the amplifier input—common-mode signals. Therefore, an amplifier circuit with a high CRR was selected to suppress these interference signals. 3. High input impedance: The contact impedance between muscle tissue and the electrodes can vary considerably. In dry areas, the contact resistance can even reach tens of thousands of ohms. Under these conditions, even with an excellent common-mode ratio, insufficient input impedance can cause output errors due to common-mode interference. Therefore, it is essential to increase the amplifier's input impedance.
Based on the above, the designed electromyography (EMG) signal acquisition circuit requires high gain, high input impedance, high common-mode rejection ratio (CMRR), low zero drift, low offset, low power consumption, and especially low 1/f noise voltage. This paper uses the INA128PA, a non-inverting parallel differential three-op-amp instrumentation amplifier from Texas Instruments' Burr-Brown series, as the core component to build the preamplifier circuit, achieving good circuit performance. The internal circuit diagram of this chip is shown in Figure 1.
Figure 1 Internal schematic diagram of INA128
The surface electromyography (EMG) signal is very weak, and the signal guided from the electrodes is mixed with strong interference signals. To avoid severe distortion caused by the signal entering the nonlinear region when the interference is strong, a two-stage amplification should be used. An instrumentation amplifier INA128 is used as the first-stage amplification, and a proportional operational amplifier is designed as the second-stage amplification.
4. Filter Design
Surface electromyography (EMG) signals typically have voltages in the millivolt range, often containing low-frequency (near DC) and high-frequency interference signals. The truly useful EMG signals are roughly between 10Hz and 500Hz. In addition, the 50Hz power frequency signal is also a significant source of interference; if not removed, it may mask the surface EMG signal. Based on these specific requirements, a dedicated filter must have DC blocking and filtering functions, and must possess a high common-mode rejection ratio and good anti-interference capabilities. A voltage-controlled voltage source type second-order low-pass filter is used for this purpose.
The 50Hz power frequency signal has a significant impact on the acquisition of surface electromyography (EMG) signals. Its frequency falls precisely in the energy concentration band of EMG signals, and its amplitude is 1-3 orders of magnitude larger than that of EMG signals, therefore it must be removed. In this design, a dual-T active filter is used to filter out the 50Hz power frequency signal, as shown in Figure 2.
The following analyzes the possible pathways that introduce power frequency interference: 1. Introduced by spatial radiation: Electromagnetic fields in space can induce currents of corresponding frequency components through electrode connections in detection equipment, connections on printed circuit boards, device pins, or the device itself, becoming noise mixed into the electromyographic signal. Electromagnetic fields in space can originate from various sources, the most critical being power frequency interference caused by grid radiation. 2. Introduced by DC power supply: In detection equipment, the DC power supply for active devices is usually obtained by transforming, rectifying, and stabilizing a power frequency AC power supply. DC regulated power supplies cannot achieve ideal filtering effects. Power frequency (or its harmonics) current, existing in the form of ripple, will be introduced into the amplifier circuit through the power supply. 3. Introduced by the subject's body: The subject's body exposed to spatial electromagnetic fields will also induce currents through the electromagnetic field. The power frequency current induced by the subject's body passes through the detection electrodes and, together with the bioelectrical signal, is added to the amplifier input, forming power frequency interference.
To address the power frequency interference introduced by the DC power supply, batteries are used to power the active devices. Using battery power not only avoids the power frequency interference problem caused by rectified and regulated power supply ripple, but also eliminates the possibility of electric shock to subjects due to leakage. Since the battery voltage is low, and using multiple batteries would result in a bulky device, a DC/DC module is used to boost the voltage and solve the chip's power supply problem.
Figure 2. Double-T active filter circuit
5 A/D Conversion
Since the sampling frequency is not high, the 8-bit serial A/D converter ADC0832 is sufficient. The ADC0832 uses a sample-data-comparator structure and employs a successive approximation method for conversion. Depending on the multiplexer's software configuration, in single-ended input mode, the input voltage to be converted is connected to one input terminal and ground; in differential input mode, the input voltage to be converted is connected to one input terminal and the other. The two inputs of the ADC0832 can be assigned as positive or negative, which can be configured by the multiplexer software. However, it should be noted that when the input voltage connected to the positive terminal is lower than the input voltage assigned as negative, the conversion result is all zeros. Control commands are transmitted via a serial data link connected to the control processor, and channel selection and input configuration are performed by software. The serial communication format allows for more functionality within the converter without increasing the package size. Furthermore, the converter and analog sensor can be placed together for serial communication with a remote control processor, instead of remotely transmitting low-level analog signals. This processing ensures that noise-free digital data is returned to the processor, avoiding interference from long-distance analog signal transmission. The hardware structure design of the entire data acquisition system is complete, as shown in circuit diagram 3:
Figure 3 System Circuit Diagram
6. Conclusion
Surface electromyography (EMG) signals are extremely weak, requiring amplification to meet the requirements of the AD acquisition unit. Furthermore, since the human body is a conductor, power frequency interference and external electric and magnetic fields can create measurement noise within the body, interfering with the detection of EMG information and severely impacting the operation of the measurement system and the accurate measurement of useful signals. Based on the characteristics of EMG signal generation and the basic requirements of acquisition technology, this paper designs a digital sensor for EMG signals, achieving good experimental results.
References
[1] Shi Shuo, Li Xijie, Wang Xu. Serial communication between electromyography signal acquisition system and computer [J]. Microcomputer Information, 2006, 1: 226-227.
[2] Surface Electromyography: Detection and Recording, Delsys Incorporated, 2002
[3] Qian Xiaojin, Zheng Rubing, Wang Chuanlin. Design of a high-performance electromyography (EMG) preamplifier. Modern Scientific Instruments, 2003, 19(3): 50-52
