Analysis of EMG Signal for Application in Prosthetic Device Motor Control (servo motor)

[Steve.J]

It's unusual to use a driven leg for EMG recording - the EMG signal is very localized and the recording electrodes are very close together and so see less common-mode interference than in a typical ECG recording. Before going to the trouble of adding this circuitry you may want to try the simpler version.  It's also easier to do things like twist the two recording electrode wires together to minimize other induced interference, which is impracticaly in ECG recordings. By the way, in addition to the two recording electrodes going to the inputs of the instrumentation amp, you will need to ground the patient (again, through a series protection resistor of 1Mohm or so) to the amplifier ground to ensure that the subject's EMG baseline does not "float" above the amplifier's common-mode input voltage range.

Regarding filter capacitors, it can be hard to find ceramic capacitors with tight tolerances to set precision filter parameters, but for EMG this should not be a big concern.  However, some care should be taken to getting the capacitor values matched for best performance.

[Jazlan]

Thank you S.E Johnson. I have noticed that the closer the two recording electrode wires are to each other, the lesser the noise in the output EMG signal.

Sir I have attached a circuit schematic (protection_circuit.jpg) showing the inputs from the electrode wires connected to 1 Mega Ohm resistors and also connected to the ground (my instrumentation amplifer ground is connected to the ground of the entire circuit) before going into the input pins of the instrumentation amplifier. Sir could u please tell me if this is the right connection in order to ensure that the subject's EMG baseline does not float above the amplifiers common mode input voltage?

For the capacitors will silver mica and metallized polycarbonate be suitable for setting frequencies of the band pass filter?

[Steve.J]

Hi Jazlan - your protection circuit won't work like that...it is just grounding the inputs your instrumentation amplifier together.  All you really need to do is tie a 1Mohm resistor to your amplifier's analog ground (if you are using two batteries to generate the postive and negative voltages, then your analog ground is the 0V node that is between the two batteries).  The other end of the resistor goes to your subject - usually somewhere near the two recording electrodes but in an less electrically active location.

Pretty much any capacitors will work if you can find them in the right value and good tolerance (+/-5% or so). The choice becomes tougher when you care a lot about factors like temperature stability, physical size, cost, etc.  Since we are dealing with low frequencies, other factors like inductance are not a concern.

[Jazlan]

Thank you sir. I have placed 100k resistors between the instrumentation amplifer input pins and the AgCl electrodes. However in my circuit I connected the 0V node analog ground from the two 9V batteries to the ground of the EMG circuit, therefore I only have one ground for the entire circuit. For the protection circuit Can I connect the 1Mohm resistor to the EMG circuit ground and connect the other end of the resistor to the subject's elbow (less electrically active location) ?

[Steve.J]

Yes - the 1Mohm resistor should go to the EMG circuit ground.  Near the elbow should be a good place for the ground electrode.

[beirut]

Hello...

I did the amplification and filtering circuit, iam waiting for the pcb to be manufactured in order to test the board.

My next step is to view the signal on a screen connected to pxi, so i will need a DAQ card and signal conditionner.

In my university the daq card is pxi-6289_M series, and the signal conditionner is SCB 68.

The data aqusition card has 18 bits resolution so i think its a good choice for this application, but i need a support concerning the connection of my sensor to scb.

[Steve.J]

Yes - the PXI-6289 will be great. You won't need all the resolution...in fact, 12 bits is sufficient for this application.  You will want to set the input mode of the DAQ card to the NRSE mode (Non-referred, single-ended). Wire your output from the EMG circuit to Channel 0 of the Analog  Inputs on the DAQ card and connect the EMG circuit ground to the AISENSE input on the DAQ card.  Alternatively, you can keep the DAQ card in the Differential mode and wire the EMG circuit ground to the negative input of on of the Analog Input 0 and the output from your circuit to the positive input of the same channel.

Now you can see the reason for the 1Mohm resistor between the patient and the EMG circuit ground - this limits any fault currents from the PC/DAQ system to a safe level while keeping the subject at the same ground potential.

[beirut]

Thank you very much Johnson...

I  attached my final schematic diagram, may be some one will benefit from it, and in case there is any problem in the design, i am waiting for feedbacks to improve my circuit.

When i test the circuit and connect it to DAQ i will send the results.

[Jazlan]

Thanks allot to Johnson for the valueble information. Also thanks to Beirut for sharing your methods.. I would like to ask:

1) For safety purposes in order to isolate the subject from the DAQ and computer, is it necessary to add an optocoupler such as the HCNR 200 between the output of the EMG amplifer circuit and inputs of the SCB-68 connector block?

2) If i want to develop two EMG circuits to measure EMG simultaneously from two different forearm muscles, can I combine the ground from these two EMG circuits and connect this ground to AISENSE input on the DAQ card? Im planning to connect the outputs from the EMG circuits to analog inputs a0 and a1.

Im using the PCI-6221 DAQ card together with SCC-68 connector block.

[Steve.J]

Hi Jazlan,

First of all, none of this kind of stuff should be done on humans without the oversight of an experienced biomedical engineer, professor, etc.

1) It is not necessary, but if done properly isolation not only creates better protection by preventing any path for DC fault currents to/from the PC/DAQ system, but it also can reduces noise from getting conducted back into the amplifier system.  The medical safety standards (like IEC 60601-1) allow for Type B inputs on medical devices (non-isolated) but these standards specify requirements for leakage currents, fault currents, and insulation testing to ensure that the subject is protected.  These requirements are usually met by series current limiting devices, AC coupling the inputs (prevents any DC fault currents), and medical grade power supplies on anything that is powered from wall power (lower AC leakage currents and high insulation integrity). For anyone connecting these systems to a PC/DAQ system, it's best to use a laptop with battery power - this way there is no possible way for AC line voltage (115/230VAC) to get back to the subject.

2) Yes - as long as the two EMG circuits share the same analog circuit ground, then you can use a single reference input AISENSE to the DAQ card. This is the advantage of the NRSE mode - you get a common analog ground reference that is not tied to the PC ground (for better noise performance).  If each amplifier has it's own ground reference then you need to use the Bipolar input mode on the DAQ card and this halves the number of channels.