INA128P EMG Circuit Not Detecting Anything

Discussion in 'The Projects Forum' started by NathanielZhu, Sep 29, 2014.

  1. NathanielZhu

    Thread Starter Member

    Dec 5, 2011
    39
    2
    Hi, I'm using a INA128P amplifier following this circuit floating around the internet:
    http://letsmakerobots.com/files/userpics/u11554/EMG_circuit_diagram.jpg

    Problem is, when I put it on my muscles, the multimeter detects around 0.1-0.3V but it changes randomly and not in response to my contraction. The electrodes (just bits of metal) all have resistance less than 1kohm (around 200 ohm).

    I recorded what I tried doing. I appreciate any help. Thanks!

     
  2. #12

    Expert

    Nov 30, 2010
    16,355
    6,852
    I see only one battery. That chip does not like inputs within 2 volts of its power voltage or its ground voltage. Have you arranged that correctly?
     
  3. blah2222

    Well-Known Member

    May 3, 2010
    554
    33
    Some problems I can see off the top of my head are that your electrodes look like dull dry washers that are very high impedance. For good signal conduction from your muscles, you're going to need a nice high conductive pathway to extract the signal. Even if you have a bit of sweat or salt water between the electrodes and your skin, it's going to be a tough bet that you'll get a decent signal as the impedance is still rather large. Also, you are going to get DC potentials that will build up at the electrode-skin interface, which can saturate your amplifier if the difference between them is large enough as they are not stable electrodes.

    In the past, I tried this out using dimes as electrodes with tape, but the connection is shaky at best and it doesn't conform well. It works, but the signal is VERY weak. If you just want to get a decent signal, I'd recommend getting some commercial ECG/EMG Ag/AgCl electrodes which are a heck of a lot more stable and predictable.

    3M Red Dot electrodes aren't bad.

    My $0.02

    EDIT***

    Also, the signal is an AC signal and using a multimeter to visualize it will be useless. An oscilloscope is your best bet to actually see it. The signal from your arm is going to be in the 10s to 1000s of microvolts and is going to need a gain stage after the front-end amplifier to get it up to something useful in the volt range. From there you can rectify the signal into a DC pulsed wave to play around with it a bit more.

    Electric fields from mains electricity are pretty much everywhere and all over your body, so without good quality matched electrodes, you will see large 50/60 Hz swamp out your small EMG signal as well. The wires connecting your electrodes to the amplifier should also be somewhat taught as their movement adds large low frequency artifact noise as well.

    Hope that helps.
     
    Last edited: Sep 29, 2014
  4. NathanielZhu

    Thread Starter Member

    Dec 5, 2011
    39
    2
    Thanks for your reply.
    The schematic says 2 batteries but I only had one at hand. I don't think 18-9V would make the device not detect....anything at all. I'm thinking it would be less sensitive but still do something. Would it???


    Thanks for responding.
    Umm, I'll consider buying proper electrodes when I can. I know the dimes/coins do work and I just want to make a setup that works (but doesn't need to work well).
    Do you see any problem with following the schematic? (I'm a hobbiest and I've never gotten a circuit more complicated than the 555 timer working).

    Aside from having 1 battery (cus I only have 1 right now and getting another where I am would be a big ripoff)
     
    Last edited: Sep 29, 2014
  5. MrChips

    Moderator

    Oct 2, 2009
    12,452
    3,371
    Measuring physiological signals is not as easy as you might assume.
    You are already making a number of mistakes as pointed out by earlier posters.

    The signals you are trying to detect are in micro-volts. You will encounter very large common mode signals. Hence your signal-to-noise ratio will be poor.

    1) You have to use two 9V batteries.
    2) You have to prepare the skin surface properly and use proper electrodes such as 3M Red Dot electrodes.
    3) You need to pay attention to the cable runs and reference cable.
    4) You need to scrap the protoboard and put the circuitry in a grounded metal case.
    5) Those two capacitors look huge and are unnecessary. What you need are two 0.1uF ceramic disc capacitors and two 10-100uF electrolyic capacitors.
    6) Following the INA, you need an additional gain stage.
     
  6. blah2222

    Well-Known Member

    May 3, 2010
    554
    33
    I'd recommend getting rid of the capacitors between the pins 1 and 8, I don't see why those are necessary at all. Read the datasheet and look at their provided application circuit.

    Like I said, even with good quality electrodes the signal is a few mVs max, and your ADC is expecting signals 0 to 5V or 3.3V. Give the signal a gain of 30 to 50 so as not to saturate the amplifier in the event of large DC offsets between the electrodes. Then place a simple RC high pass filter fed into a buffer op-amp to get rid of any DC (as the EMG signal is AC). Then put in an op amp gain stage of about 100 to boost it up to the level the ADC can use. Finish that off with either a passive or active low pass filter before the ADC to eliminate aliasing effects (limiting the signal's bandwidth to less than half the ADCs sampling frequency).

    I cannot stress the importance of the third electrode placed pretty much anywhere on the arm. It is used to bias the system ground to your body which is filled with mains (50/60 Hz) noise. It basically cuts that out quite a bit, leaving you with a fairly nice signal.

    As MrChips touched upon, wiring this up and expecting it to work like a regular circuit is not going to happen. It would be good to read a bit more about EMG.

    EDIT***

    ADC stands for Analog to Digital Converter. It converts the analog AC EMG signal to a digital representation that can be transferred or stored.

    I'd recommend checking out the following YouTube channel to get a bit more accustomed to biosignals/amplifier design, if you are interested in pursuing this further.

    https://www.youtube.com/user/bratwursthaggisgeek
     
    Last edited: Sep 29, 2014
Loading...