Overview

In order to construct an EMG properly, many design parameters must be considered. The device must be able to detect the small signals arising from muscular contraction and amplify them without also amplifying ambient noise which may exist in the recording environment and which is likely to be of approximately the same magnitude as the signals of interest. Another critical consideration when designing an EMG for use on human beings is that the subject must always remain galvanically isolated from any external ground. Thus, if almost any external measuring or recording device is used to analyze the signal, the signal must be transmitted over some non-conductive media.

Specifics

To insure that our subject was never connected to any external ground, we chose to transmit our signal over the FM band from one collection and amplification circuit to a second circuit where the signal was processed and output to an oscilloscope. This meant that our EMG would consist of two completely discrete circuits, both of which are described below.

The collection circuit was comprised of a number of different stages and integrated circuits. To insure that our EMG was able to detect and isolate the signals in which we were interested we chose to use an INA121 instrumentation amplifier from Burr-Brown. This amplifier took as input three electrodes attached to the subject. Two of the electrodes were attached along the muscle belly. The voltage difference between these two electrodes was the EMG signal and was amplified by the INA121 by a factor of 1000 in order to produce an output in the volt range. Any voltage that appeared on both electrodes would not amplified by the INA121, which had a CMRR of 106 dB at 1000 gain. In this manner, ambient noise was removed from our signal. The third electrode was connected to an inactive site on the subject and provided a reference voltage for the amplifier. A schematic of this portion of the circuit follows:

To transmit the signal, a BA1404 FM transmitter was used. A schematic of this portion of the circuit follows:

However, since the signal was being transmitted over the FM band and since FM does not transmit low frequencies (such as some of the frequencies in the EMG spectrum) well, voltage levels output from the INA121 were translated into pulses of varying frequencies using an LM331 Voltage-to-Frequency Converter from National Semiconductor. This converter was set so that an input signal of 0 volts corresponded to frequency output in the kilohertz range, thereby assuring that the limitation of the FM band would not fail to transmit the EMG signal. A schematic of the VFC portion of the circuit follows:

Since this circuit could not be connected to ground, it also had to contain its own power supply. To accomplish this, a voltage regulator was built which both served to isolate the subject from ground and provide an appropriate voltage range so that that amplifier could respond to both positive and negative signals produced by the muscle. A schematic of this regulator follows:

The input to the collection and output circuit was a standard FM receiver. Once the signal had been output from the radio, it was passed through two operational amplifiers which acted first as a 60 Hz notch filter and then as a non-inverting amplifier with a gain of approximately 4. The 60 Hz filter was implemented in order to remove some of the noise, which might be introduced as a function of FM transmission. The purpose of the non-inverting amplifier was to insure that the signal would always be of sufficient voltage to trigger the Frequency-to-Voltage Converter. A schematic of these two op-amps follows:

After this stage, the signal was passed to the Frequency-to-Voltage Converter (LM331) which translated the frequencies produced in the collection circuit back into voltage levels for further processing.

The last stages through which the signal passed consisted of three operational amplifiers which filtered the signal, rectified the filtered signal, and integrated the rectified signal, respectively. The filter chosen was low-pass filter with a time constant of approximately 500 Hz. This frequency was well above the expected range for EMG and served to filter extraneous high-frequency noise. The signal could be output after each of these three stages. A schematic for these amplifiers follows:

Since the collection circuit was not connected to the subject, it could be powered from the power source present on the prototype board. No voltage regulation component was therefore necessary for this circuit.