Biomedical User Group Discussions

cancel
Showing results for 
Search instead for 
Did you mean: 

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

Yes ofcourse... i can share the external circuit for the amplifier and the filter..

With respect to the amplifier gain, it can be changed according to the value of Rg,

so what i will do is that i will put a reohstat and change the gain until i find the output value a good value to be measured and analysed.

The Filter i designed is based on butterworth filter design, it is formed of 2 filters, the first one is low pass, and the other is high pass filter,

to make band pass filter f1=20Hz and f2=500Hz

I am not concerned to safety, since am not suppling my boards from 220v, am suppling from external battery source.

0 Kudos
Message 11 of 64
(1,982 Views)

Even though you are battery powered, you should be careful about applying electrodes to a person with no current limiting devices. Even with 5V supplies, you should put series resistors (maybe 100Kohm) on each electrode input to your instrumentation amplifier in case there is a short to the supply voltage, limiting any fault current to 50uA.

Why are you not running the AD620 on +/- 5V, the same as your filter circuits? Your EMG signal will be clipped at zero (only positive signal swing at the output).

You are setting gain the first stage gain to about 225.  I think you will find this to be too aggressive, since any small differences in DC electrode offset potential will be also amplified by this factor. So if there is a difference of 50mV in the half-cell potential of your electrodes (very common) then your AD620 will attempt to make over 11V at the output - no good. Lower this gain and raise the gain later in the circuit and you should be better off.

Finally, why are you bringing the Reference input from the AD620 out to a connector labeled "REF-ELECTRODE"?  This input on the AD620 is normally used to bias the output up/down to better match the input range of an A/D converter, for example. I think you just need to ground this, and the same ground will be the reference for your output signal on J1 pin 1.

Visit the NI Biomedical User Group at:
www.ni.com/biomedusers
0 Kudos
Message 12 of 64
(1,982 Views)

I just did a little simulation on your circuit...a couple of suggestions:  1) try running your amplifier on two 9V batteries in series to give your op-amps more headroom - +9V to positive supply terminal and -9V to negative supply terminal with the center of the two batteries as your analog ground (0V).  This gives you 18V of swing for the instrumentation amp.  2) run your AD620 at a gain of 20 using 2.61K resistor and set your variable gain op-amp to a gain of 50 (input resistor 1K and feedback resistor 50K or similar). You may not need this much gain depending on your signals and the input range of your ADC so you can adjust this as needed.  Also forgot to mention that those 100K series resistors for you inputs will also help protect the AD620 from harm from static electricity, in addition to protecting the subject!

Visit the NI Biomedical User Group at:
www.ni.com/biomedusers
0 Kudos
Message 13 of 64
(1,982 Views)

Thanks for your comments Johnson...

I looked back to my design and did some simulation in proteus software, and come with the following new conclusions:

1- I will supply the intrumentation amplifier with +5v and -5V in order not to eleminate the negative part of the signal, and i will make the first stage gain to 5 by adjusting the value of the resistor to 12.35K.

2- I will set the reference pin to 0V, and after the second stage amplifier i will add an offset voltage to the signal in order to enter it to the ADC, with a gain of 200, hence i will have a gain of 1000 at the input of ADC.

I have something that i didnt understand, you said that increasing the gain of the first stage is not good since any small dc offset from the electrodes will be multiplied by the same value, but if this offset is not eliminated at all it will also be amplified in the second stage amlplifier so the same problem present right?

0 Kudos
Message 14 of 64
(1,982 Views)

Your high-pass filter will block the DC offset that is present after the first stage amplifier - it will only let the AC signal pass through to your second gain stage. You can play with the two gains until you get the best performance.

Visit the NI Biomedical User Group at:
www.ni.com/biomedusers
0 Kudos
Message 15 of 64
(1,982 Views)
Highlighted

Yes thats i was thinking about, that the filters should block such offset.

I attached the new design i did, and please if you have comments and improvements let me know..

Thanks a lot Johnson

0 Kudos
Message 16 of 64
(1,982 Views)

I have also constructed a similar EMG signal conditioning circuit. Ive attached the schematic and oscilloscope outputs for the band pass filter stage, precision rectifier stage and smoothing low pass filter stage. However your high pass and low pass filters are better 2nd order filters than mine. Which muscles are you measuring EMG from? I measured these signals from my biceps brachii. Im currently constructing a circuit for measuring EMG from flexor digitorum superficialis and extensor carpi radialis longus forearm muscles

In this circuit the gain of the instrumentation amplifier is only around 20 to prevent small DC offsets from being amplified. Thats why I further amplified the signal at the low pass filter stage. The high pass filter has unity gain but the low pass filter has a gain of 160. The low pass filter not only filters the signal but amplifies it as well. After high pass and low pass filtering, i use this half wave precision rectifier to eliminate the negative portion of the signal

I have also made the similar mistake of not placing resistors at the instrumentation amplifier inputs to protect the user. Where will you be connecting the outputs of the EMG amplifier? There are some conference papers which use optocouplers (such as National 4N25) at the end of the EMG amplifer circuit before the output of the EMG circuit is sent to a Labview DAQ through SCC-68 connector block and PCI-6221.

In chapter 3 of the masters thesis by Saksit page 25, there is a driven right leg circuit diagram with shielding. He used RCA to stereo wires to connect the AgCl electrodes to the inputs of the instrumentation amplifier.

http://medusa.sdsu.edu/Robotics/Neuromuscular/Theses/Saksit/Saksit.htm

This circuit improves the signal to noise ratio.

0 Kudos
Message 17 of 64
(1,982 Views)

beirut what type of capacitors are you using in your filter design? Film (polyester or polypropelene) or ceramic/monolithic type? I used ceramic capacitors in my circuit. Im working on a similar project also. My email is ahmad_jazlan06@yahoo.com if you would like to contact me.

0 Kudos
Message 18 of 64
(1,982 Views)

Hello Jazlan..

I used ceramic capacitors as you.

did you take into consideration the driven right leg circuit?

I want to ask another question concerning the range of voltage which is measured between the 2 electrodes connected to the muscle.

0 Kudos
Message 19 of 64
(1,982 Views)

Hello Beirut. =)

In the circuit schematic attached previously I didnt use the driven right leg circuit but now I am after my teacher recommended it. I have attached the schematic of the driven right leg circuit from Saksit's thesis. He used an INA 2128 instrumentation amplifer but in your case i think the AD620 u are using is fine.

The value of the EMG voltage before going into the instrumentation amplifer is around 1-5milivolts i think. Depending on the muscle type and strength of muscle contraction. Biceps and trapezius are stronger than forearm muscles.

If ur interested in constructing a driven right leg circuit with shielding:

Basically in Figure 1, I sliced off the red and white banana plugs and replaced them with crocodile clips (to clip onto the AgCl electrodes). When we remove the insulation of the left and right wires, we can see inside the left wire there is one smaller insulated wire (white or red insulation) and another shield wire with no insulation. Similarly we can see inside the right wire, there is one smaller insulated wire (white or red insulation) and another shield wire. Strip off the insulation of both the white and red smaller wires and solder these wires to crocodile clips to be connected to AgCl electrodes. The shield wire is left alone but make sure it does not accidentally touch the crocodile clips or other parts.

The green stereo audio jack of the RCA to stereo audio cable is connected to a female connector (I dont have a picture at the moment). This female connector has 3 parts which can be soldered (positive,negative and shield). The positive and negative parts of the female connector are wired to inputs 2 and 3 of the instrumentation amplifier. In Figure 2, at the 'phone plug' 1 and 3 represent the positive and negative parts of the female connector whereas 2 represents the shield. 1 and 3 are connected to the instrumentation amplifier, 3 is connected to the driven right leg part which begins with the inverting pin of U2A.

0 Kudos
Message 20 of 64
(1,982 Views)