Copy the page URI to the clipboard
Poulton, Adrian
(2010).
URL: http://www.ot-forum.de/trade_shows_congresses/orth...
Abstract
SUMMARY:
Electromagnetic compatibility of myoelectrode amplifiers for prosthetic control is important for safe operation in electrically noisy conditions. Factors affecting susceptibility to interference (impedance, common mode rejection ratio (CMRR) and isolation) were studied using a commercial amplifier.
INTRODUCTION:
Electrical interference can enter the system either directly between the active electrodes as a differential signal, or via the common electrode as a common mode signal. The direct path is susceptible to interference due to the high gain of the amplifier. Shielding and electrode geometry design can help reduce the effects. A notch filter is generally used to reduce greatly the amplifier gain at mains frequency, though harmonics are still passed. Common mode interference is mitigated by the normally very high CMRR of the amplifier; however, common mode signals can also be converted to differential signals through imbalances in electrode impedances (Winter and Webster, 1983; Scott and Lovely, 1986). Common mode interference mainly enters via the power and output leads of a myoelectrode amplifier, so effective isolation should reduce the effect.
METHODS:
Two Otto Bock 13E125 active myoelectrodes were used for the investigations. Sinusoidal signals of different frequencies were applied in differential and common mode configurations and the output measured. The myoelectrode and measuring equipment were all battery powered to avoid conductive mains interference.
Connection to the electrode contacts was through an assembly of spring-loaded platinum contacts, and the myoelectrode was shielded from the external environment in a die-cast box. Measurements were made with different balanced and unbalanced input impedances to represent the skin interface. The measurements were made with the myoelectrode alone and also in combination with an isolation amplifier (Burr-Brown ISO124), and a d.c.-d.c. converter (Murata MEA1D0505SC) to provide isolated power.
RESULTS:
The d.c. input impedance and intrinsic CMRR of the myoelectrode amplifier were extremely high by design (>40 M? and >90 dB). The response to differential signals was greatest at 200 Hz to 1 kHz, with a sharp notch at 50 Hz. The CMRR was significantly reduced for a.c. signals if the electrode impedances were unbalanced. As the amplifier gain was non-linear and frequency dependent, a protocol was followed where the imbalance in electrode impedances was adjusted to give the same output as a known differential input. Thus the effect of impedance imbalance could be separated from the intrinsic CMRR of the amplifier. The results obtained fitted the Winter-Webster model with input capacitances of 400 pF. The centre earth electrode was found to be strongly coupled to the 0V lead at a.c. via a capacitance of 1 ?F. Isolating the amplifier improved the rejection of common mode signals introduced via the 0V lead. There was a 20 dB increase in CMRR with the isolation components used.
CONCLUSION:
Common mode interference can enter via the leads of a myoelectrode amplifier. Even if the intrinsic CMRR of the amplifier is very high, common mode is converted to an interfering differential signal if the electrode impedances are unbalanced. This is very likely to be the case, due to differences in skin contact. It has been demonstrated that the common mode route can be blocked by an isolating amplifier and dc-dc converter, though at the expense of extra complexity and the need to power these components. Practical implementations may come from developments in low power circuitry for applications such as wireless sensor networks.
REFERENCES:
Winter B. and Webster J. 'Reduction of interference due to common mode voltage in biopotential amplifiers', IEEE Transactions on Biomedical Engineering, BME-30 (1), 58-62, 1983.
Scott R.N. and Lovely D.F., 'Amplifier input impedances for myoelectric control', Medical & Biological Engineering & Computing, 24, 527-530, 1986.